Datasets:

ccdv
/

arxiv-summarization

Dataset card Data Studio Files Files and versions Community

Subset (2)

section · 216k rows

Split (3)

train · 203k rows

article stringlengths 0 745k	abstract stringlengths 0 92.7k
the smale - hirsch theory @xcite reduces the problem of enumerating regular homotopy classes of immersions to homotopy theory . however , the problem of determining the regular homotopy class of a given immersion remains nontrivial . here we address this problem in the case of immersions of @xmath0-manifolds into @xmath12 having trivial normal bundles . we concentrate on immersions with trivial normal bundles , since we want to apply our results to the following question : _ which regular homotopy classes of immersions of an oriented @xmath0-manifold into @xmath12 contain embeddings ? _ note that embeddings of oriented @xmath0-manifolds into @xmath12 have trivial normal bundles . _ history . _ the problem of reading off the regular homotopy class of an immersion from its picture was posed by smale @xcite . for the particular case of immersions of the @xmath13-sphere @xmath14 into @xmath15 , it had already been shown by whitney @xcite that the double points of an immersion @xmath16 determine its regular homotopy class . this result has been extended by ekholm @xcite to immersions @xmath17 and @xmath18 on the other hand , a result by hughes and melvin @xcite suggested that no further extension was possible . namely they have shown that there are embeddings @xmath19 not regularly homotopic to the standard one . therefore the multiple points do not determine the regular homotopy class of an immersion . still the original problem posed by smale has been answered by ekholm and szcs @xcite for immersions @xmath20 by considering `` singular seifert surfaces '' bounded by the immersions . immersions of closed oriented @xmath0-manifolds into @xmath12 were investigated by wu @xcite . using the homotopy theoretical reduction provided by the smale - hirsch theory he has shown that the regular homotopy classes of immersions of a closed oriented @xmath0-manifold @xmath21 into @xmath12 can be characterised by two invariants as follows . * the first one is a second cohomology class of @xmath21 : `` half of the normal euler class '' . * the second invariant is an integer ( for immersions with trivial normal bundles ) . the geometric meaning of these invariants was not very clear from wu s work . _ now we can list our results : _ 1 . we shall give a geometric description of the first invariant in the first part of section [ sect : c ] . we call it `` the wu invariant '' . note that for an embedding @xmath22 the wu invariant is either zero or an element of order two in the second cohomology group @xmath23 of @xmath24 2 . we show that given any second cohomology class in @xmath23 , which is zero or has order two , there exists an embedding @xmath25 of the oriented @xmath0-manifold @xmath21 with wu invariant equal to the given cohomology class ( theorem [ thm : c ] ) . we define an action of the group @xmath26 $ ] of regular homotopy classes of immersions of @xmath27 into @xmath12 on the set @xmath28_0 $ ] of regular homotopy classes of immersions of the @xmath0-manifold @xmath21 into @xmath12 with trivial normal bundles by taking connected sums of immersions . in section [ sect : sharp ] , we show that this action is effective and its orbits coincide with the regular homotopy classes of immersions having the same wu invariant . [ item:4 ] in section [ sect : main ] , we express the second invariant ( which is an integer ) through singularities of any singular seifert surface . actually we prove two formulae , analogous to those proved by ekholm and szcs @xcite for immersions of @xmath27 into @xmath12 . 5 . in section [ sect : t3 ] , we prove a few rather surprising corollaries concerning the question mentioned above about the regular homotopy classes containing embeddings . although these corollaries hold more generally , we formulate them for the simplest manifold for which these surprising phenomena occur , namely for the @xmath0-dimensional torus @xmath5 : + * corollary [ cor : t3sz ] . * _ there exist embeddings @xmath3 and @xmath4 of the @xmath0-torus @xmath5 with the following properties : @xmath6 there is an immersion @xmath9 such that @xmath8 is regularly homotopic to the connected sum @xmath29 , but @xmath10 this immersion @xmath11 can not be chosen from the regular homotopy class of an embedding . _ + * corollary [ cor : t3 ] . * _ there exists an immersion @xmath30 with the following properties : @xmath6 @xmath11 is not regularly homotopic to an embedding @xmath31 , but @xmath10 for any embedding @xmath32 , the connected sum @xmath33 is again regularly homotopic to an embedding @xmath34 . _ + in other words , the double points of the immersion @xmath11 can not be eliminated by regular homotopy , but they can be eliminated after being taken connected sum with any embedding of the @xmath0-torus . 6 . in section [ sect : cusp ] , we give some corollaries concerning the number of cusps a singular seifert surface can have . we will use the following notation throughout the paper . by @xmath35 $ ] we denote the set of regular homotopy classes of immersions of a manifold @xmath36 into a manifold @xmath37 and by @xmath38 $ ] the subset of @xmath35 $ ] consisting of all regular homotopy classes containing an embedding . by @xmath35_{\chi}$ ] we denote the subset of @xmath35 $ ] consisting of all regular homotopy classes of immersions with normal euler class @xmath39 . in particular for an oriented @xmath0-manifold @xmath21 by @xmath28_0 $ ] we denote the set of regular homotopy classes of immersions @xmath40 with trivial normal bundles . by @xmath21 we will always denote a smooth , closed , oriented @xmath0-manifold . we denote by @xmath41 the finite subgroup of @xmath23 formed by zero and all elements of order two . the punctured manifold of @xmath21 will be denoted by @xmath42 ; i.e. @xmath43 ( @xmath44 is a @xmath0-disk ) and the @xmath45-skeleton of a manifold @xmath36 by @xmath46 . throughout the paper , manifolds and maps are of class @xmath47 . the symbol `` @xmath48 '' denotes an appropriate isomorphism between algebraic objects ; `` @xmath49 '' and `` @xmath50 '' mean `` homotopic '' and `` regularly homotopic '' respectively . hughes @xcite has shown that @xmath51 $ ] has a group structure under connected sum and the smale invariant @xmath52 \to \pi_n(v_{n , n})\ ] ] actually gives a group isomorphism , where @xmath53 denotes the stiefel manifold of @xmath13 frames in the @xmath37-space . hughes and melvin @xcite investigated the problem : which regular homotopy classes contain embeddings in the set @xmath54 $ ] ? their result for the case of embeddings of @xmath27 into @xmath12 can be summarised by the following diagram : @xmath55 & \stackrel{\approx}{\longrightarrow } & \pi _ 3 ( v_{5,3})\approx { \mathbf{z}}\\ & \cup & & \cup \\ & { \mathop{\mathrm{emb}}\nolimits}[s^3 , { \mathbf{r}}^5 ] & \stackrel{\approx}{\longrightarrow } & 24{\mathbf{z}}\\ & f & \longmapsto & 3\sigma ( v^4_f)/2 , \end{array}\ ] ] where @xmath56 is the smale invariant and @xmath57 is the signature of an oriented seifert surface @xmath58 for @xmath59 . here , `` @xmath48 '' denotes a group isomorphism . their result has been extended by the third author to @xmath60-homology 3-spheres @xcite . trying to extend it further was the starting point of the present paper . as we mentioned in the introduction , * this result shows the impossibility of determining the regular homotopy class of an immersion by its multiple points , and * it has been shown in @xcite that hughes - melvin s formula stating that the smale invariant of an embedding is equal to @xmath61 times the signature of any seifert surface bounded by the embedding can be extended to immersions if we allow `` singular seifert surfaces '' . since we are going to generalise the result ( 2 ) to immersions of arbitrary oriented @xmath0-manifolds with trivial normal bundles , we recall it in detail . let @xmath62 be an immersion and @xmath63 an arbitrary compact oriented 4-manifold with @xmath64 . the map @xmath59 extends to a generic map @xmath65 which has no singular points near the boundary . this map @xmath66 has isolated cusps , each of which has a sign . let us denote by @xmath67 their algebraic number . let us denote by @xmath68 the smale invariant of @xmath59 . the first formula for the smale invariant given by @xcite is the following : [ thm2.1 ] @xmath69 before giving the second formula for the smale invariant we recall some definitions from @xcite . let @xmath63 be as above and let @xmath70 be a generic map nonsingular near the boundary , such that @xmath71 and @xmath72 in @xmath12 . here @xmath73 is the upper half space in @xmath74 and @xmath75 . the double point set @xmath76 of the map @xmath77 is 2-dimensional , and on the boundary of @xmath78 lies the image @xmath79 of the singular points of @xmath80 , which is 1-dimensional . furthermore , triple points appear as isolated points , each of which has a sign . we denote the algebraic number of triple points of @xmath77 by @xmath81 . [ dfn : l ] let @xmath82 be the set of double points of @xmath59 and @xmath83 its image . then @xmath83 is a closed @xmath84-manifold in @xmath12 . since @xmath27 is @xmath2-connected , @xmath59 has a unique ( up to homotopy ) normal framing ( @xmath85 ) . let @xmath86 be the manifold obtained from @xmath83 by shifting slightly along the vector field , @xmath87 for @xmath88 . define @xmath89 to be the linking number in @xmath12 of @xmath90 and @xmath86 . [ dfn : l ] let @xmath91 be a copy of @xmath79 shifted slightly along the outward normal vector field of @xmath79 in ( the closure of ) @xmath92 . define the linking number @xmath93 of the map @xmath80 to be the linking number of @xmath91 and @xmath94 in @xmath73 . [ thm : e - s ] the smale invariant of @xmath59 is given by the formula : @xmath95 as is stated in ( * ? ? ? * remark 5 ) , the above definition of @xmath96 makes sense for immersions of arbitrary closed oriented @xmath0-manifolds . furthermore , the definition of @xmath97 can be extended to the case of framed immersions immersions with trivial normal bundles and endowed with normal framings of closed oriented @xmath0-manifolds , which are not necessarily @xmath27 . we review this extension here , since it will be necessary later . [ dfn : lnu ] let @xmath98 be a framed immersion with a normal framing @xmath99 , of a closed oriented @xmath0-manifold @xmath21 into @xmath12 . let @xmath100 be the set of double points of @xmath98 and @xmath101 its image . let @xmath102 and @xmath103 be the two linearly independent normal vectors of @xmath98 determined by @xmath99 . define the vector field @xmath104 along @xmath101 by @xmath105 , where @xmath106 . then , define @xmath107 to be the linking number in @xmath12 of @xmath108 and @xmath101 pushed off along @xmath104 out of @xmath108 . we will show later in the proof of theorem [ thm : b ] that @xmath107 actually does not depend on the choice of the normal framing @xmath99 for an immersion @xmath109 with trivial normal bundle of any closed oriented @xmath0-manifold @xmath21 ( see remark [ rmk : lindep ] ) . in this section , we first recall the result of wu @xcite on the enumeration of regular homotopy classes of immersions of an oriented @xmath0-manifold @xmath21 into @xmath12 . then we study the structure of @xmath28_0 $ ] the set of regular homotopy classes of immersions with trivial normal bundles . we look especially into a certain second cohomology class which characterises ( the regular homotopy class of ) an immersion on ( the regular neighbourhood of ) the @xmath2-skeleton of @xmath21 . _ we fix a trivialisation @xmath110 of @xmath111 once and for all . _ then the smale - hirsch theory provides a bijection @xmath112\stackrel{\approx}{\to}[m^3 , v_{5,3}]$ ] . based on this bijection , wu @xcite has shown the following . [ theorem 2 in @xcite , see also ( * ? ? ? * theorem 3 ) ] [ thm : wu ] the normal euler class @xmath113 for an immersion @xmath109 is of the form @xmath114 for some @xmath115 , and for any @xmath115 , there is an @xmath98 such that @xmath116 . furthermore , @xmath117_{\chi } \approx \coprod_{c\in h^2(m^3;{\mathbf{z}})\text { with } 2c=\chi } h^3(m^3;{\mathbf{z}})/(4c\smile h^1(m^3;{\mathbf{z}})),\ ] ] where @xmath112_{\chi}$ ] is the set of regular homotopy classes of immersions with normal euler class @xmath118 , @xmath119 represents the cup product and @xmath48 denotes a bijection . [ rmk : c ] let us consider the @xmath120-bundle @xmath121 and choose the generator @xmath122 for which the euler class of this bundle coincides with @xmath123 . let @xmath124 be the map associated with an immersion @xmath109 ( and the fixed trivialisation @xmath125 ) . the first part of the above theorem is based on the fact that the normal euler class @xmath113 is the pullback @xmath126 for the generator @xmath127 for @xmath128 . hence , the class @xmath115 which appears in the second part of the theorem is nothing but the pullback @xmath129 of the generator @xmath127 of @xmath128 . recall that @xmath41 is the finite set @xmath130 . hereafter we shall concentrate on immersions with trivial normal bundles , i.e. on @xmath112_0 $ ] . by theorem [ thm : wu ] this set can be identified with @xmath131 ( this identification depends on the trivialisation of @xmath111 ) . note that theorem [ thm : wu ] can also be applied to the nonclosed @xmath0-manifold @xmath42 and gives the following bijection : @xmath132_0 \approx \gamma_2(m^3).\ ] ] note that elements of order two in @xmath23 can be naturally identified with those in @xmath133 . [ dfn : c ] the projection @xmath134_0 \to \gamma_2(m^3)$ ] is called _ the wu invariant _ of the immersion of the _ parallelised _ @xmath0-manifold with trivial normal bundle . it can be described in two equivalent ways : * @xmath135 or * @xmath136 if we use the identification @xmath137_0 \gamma_2(m^3).$ ] next we shall give a more geometric description of the wu invariant @xmath138 . a normal trivialisation @xmath99 of @xmath98 ( together with the tangent trivialisation and the differential of @xmath98 ) defines a map @xmath139 , i.e. an element @xmath140 in @xmath141 . if we change @xmath99 by an element @xmath142 = h^1(m^3;{\mathbf{z}})$ ] , then the class @xmath140 changes by @xmath143 , where @xmath144 is the mod @xmath2 reduction map @xmath145 . hence the coset of @xmath140 in @xmath146 does not depend on @xmath99 . the cokernel of @xmath144 can be identified with @xmath41 by the canonical map induced by the bockstein homomorphism ( see below ) . under this identification , the coset of @xmath140 corresponds to the wu invariant @xmath147 . thus the wu invariant describes the immersion on the @xmath84-skeleton ( and then actually also on the @xmath2-skeleton , since @xmath148 ) , if a trivialisation of the tangent bundle is fixed . on the other hand for any immersion @xmath149 there is a trivialisation of the tangent bundle such that the wu invariant of the immersion under the given tangent bundle trivialisation is zero . this geometric description will follow from the lemmas below . let @xmath150 be an immersion with trivial normal bundle . choosing a normal framing @xmath99 for @xmath98 , we naturally obtain a map @xmath151 . [ lem : cx ] let @xmath152 be the generator of @xmath153 and @xmath154 the induced homomorphism @xmath155 . then @xmath156 does not depend on the choice of the normal framing @xmath99 , and is equal to @xmath157 . from the gysin exact sequence of the @xmath158-bundle @xmath159 , it is easy to see that @xmath160 . hence @xmath156 is equal to @xmath157 and does not depend on the choice of the normal framing @xmath99 . consider the following commutative diagram : @xmath161 where each horizontal line is a part of the cohomology exact sequence associated with the coefficient exact sequence @xmath162 , each vertical line is the homomorphism induced by @xmath163 , and each @xmath164 is the bockstein homomorphism . [ lem : cy ] let @xmath165 be the generator of @xmath166 . then @xmath167 does not depend on the choice of the normal framing @xmath99 , and is equal to @xmath157 . since @xmath168 and @xmath153 in the diagram above , we see that @xmath169 . hence @xmath170 . then the result follows directly from lemma [ lem : cx ] . lemma [ lem : cy ] implies that the coset of the class @xmath171 modulo @xmath172 does not depend on the choice of the normal framing @xmath173 note also that @xmath174 is isomorphic to @xmath175 by the exact sequence above . thus , under this isomorphism , we have @xmath176 $ ] , where the bracket means the coset modulo the image of @xmath177 since the class @xmath178 mentioned above is nothing but the homomorphism @xmath179 , it is clear that @xmath180 where we consider the second @xmath165 as the identity homomorphism @xmath181 . thus , we have also the description @xmath182 $ ] . henceforth we often identify @xmath41 with @xmath175 and consider @xmath157 as an element of @xmath175 . [ rmk : spin ] since we have fixed a trivialisation of @xmath111 , the set @xmath183 can be identified with the set of homotopy classes of trivialisations of @xmath184 , where @xmath185 denotes the @xmath2-skeleton of @xmath21 , i.e. with the set of _ spin structures _ of @xmath21 . the above observations show that an immersion of @xmath21 into @xmath12 with trivial normal bundle determines a spin structure of @xmath21 up to elements of @xmath172 . now we can state the main result of this section , which claims that each @xmath186-component in @xmath28_0 \approx { \mathbf{z}}\amalg{\mathbf{z}}\amalg\cdots\amalg{\mathbf{z}}= \gamma_2(m^3 ) \times { \mathbf{z}}$ ] contains an embedding . [ thm : c ] for every element @xmath187 , there exists an embedding @xmath188 with the wu invariant @xmath189 . \(a ) first assume that @xmath190 $ ] , where @xmath191 . consider the spin structure of @xmath21 corresponding to @xmath191 under the identification in remark [ rmk : spin ] . then by a result of kaplan @xcite , there exists a framed 4-manifold @xmath192 such that * @xmath193 , * the framing of @xmath192 restricted to the @xmath2-skeleton @xmath185 of @xmath21 coincides with the given spin structure of @xmath21 , * @xmath192 has a special handlebody decomposition , consisting of one 0-handle and some 2-handles attached to the 0-handle simultaneously . since @xmath192 is a nonclosed spin 4-manifold , it is parallelisable and so it can be immersed into @xmath12 . furthermore , since @xmath192 has a @xmath2-complex as its spine , its immersion into @xmath12 can be deformed into an embedding . take an embedding @xmath194 and set @xmath195 . now it suffices to show that @xmath196 $ ] . taking the normal framing @xmath99 for @xmath98 corresponding to the normal vector field of @xmath197 , we see that the map @xmath151 restricted to the @xmath2-skeleton can be written as the following composition : @xmath198 where @xmath199 is the map determined by the differential of @xmath200 using the given framing on @xmath192 . thus @xmath201 , since @xmath202 . \(b ) in the case when @xmath203 can not be written as @xmath204 $ ] with @xmath205 , put @xmath206 $ ] with @xmath207 . as in ( a ) , consider the spin structure of @xmath21 corresponding to @xmath208 , and take an embedding @xmath209 of a @xmath210-manifold @xmath192 with @xmath211 such that @xmath192 has the same properties as in ( a ) . set @xmath212 . in order to show that @xmath213 $ ] , recall that we have fixed a trivialisation @xmath110 of @xmath111 to obtain the map @xmath214_0\to\gamma_2(m^3).$ ] if we take another trivialisation @xmath215 of @xmath111 , then we obtain another map @xmath216_0\to\gamma_2(m^3)\ ] ] and with respect to this new trivialisation @xmath215 , we have a new map @xmath217 and a new induced homomorphism @xmath218 which satisfies lemma [ lem : cy ] . furthermore , if we choose @xmath215 so that @xmath219 corresponds to @xmath220 as a spin structure , then we have @xmath221 and hence @xmath222 since the restriction of @xmath223 to the @xmath2-skeleton @xmath185 can be decomposed into the same composition @xmath224 as in ( a ) , by the same reason as in ( a ) we have @xmath225 and hence @xmath226 . thus @xmath227 is a required embedding . in this section we define an effective action of the group @xmath26 \approx { \mathbf{z}}$ ] on the set @xmath112_0 $ ] and show that its orbits coincide with the sets of regular homotopy classes having the same wu invariant @xmath228 . let @xmath21 be a closed oriented @xmath0-manifold . let @xmath44 be a @xmath0-disk , which from now on we will often identify with the northern hemisphere of the @xmath0-sphere @xmath27 . fix an inclusion @xmath229 , and put @xmath230 . suppose @xmath231 is an immersion with trivial normal bundle such that @xmath232 coincides with the standard embedding @xmath233 restricted to the northern hemisphere . for an immersion @xmath62 , we may assume that @xmath59 restricted to the southern hemisphere is standard . then consider the map @xmath234 & \longrightarrow & { \mathop{\mathrm{imm}}\nolimits}[m^3 , { \mathbf{r}}^5]\\ & f & \longmapsto & f_0 \sharp f , \\ \end{array}\ ] ] where @xmath235 and @xmath236 . the normal bundle of @xmath3 is trivial and if @xmath3 is modified on @xmath44 , then its normal bundle does not change . furthermore , @xmath237 also does not change under this operation , since the invariant @xmath228 is determined by the immersion restricted to a neighbourhood of the @xmath2-skeleton of @xmath21 . therefore , if we define , for @xmath115 with @xmath238 , @xmath117_0^{c}:=\{f\in{\mathop{\mathrm{imm}}\nolimits}[m^3 , { \mathbf{r}}^5]_0\ , \| \ , c(f)=c\},\ ] ] then we can in fact define the map @xmath239 \longrightarrow { \mathop{\mathrm{imm}}\nolimits}[m^3 , { \mathbf{r}}^5]_0^{c(f_0)}.\ ] ] the following proposition is an analogue of proposition 3.1 in @xcite . [ prop : bij ] the map @xmath239 \longrightarrow { \mathop{\mathrm{imm}}\nolimits}[m^3 , { \mathbf{r}}^5]_0^{c(f_0)}\ ] ] is a bijection . as it has been mentioned in the paragraph just after remark [ rmk : c ] , we have @xmath132_0^{c(f_0)}\approx h^3({{m}_{\circ}}^3;{\mathbf{z}})= { 0}.\ ] ] this means that @xmath240 is surjective . let us prove the injectivity . for two immersions @xmath59 and @xmath241 , we have @xmath242 note that @xmath243 and @xmath244 are homotopic on the @xmath2-skeleton of @xmath21 . let @xmath245 be the difference @xmath0-cochain between @xmath243 and @xmath246 , that is , the 3-cochain which assigns @xmath247 to @xmath44 considered as a @xmath0-cell , and @xmath248 to the other 3-cells . then we have @xmath249 using the fact that @xmath250 and the following lemma . two maps @xmath251 and @xmath252 are homotopic if and only if * @xmath253 , and * there exist a @xmath84-cocycle @xmath254 and a @xmath2-cochain @xmath255 such that @xmath256 note that @xmath127 is the generator of @xmath257 and that @xmath258 ( see remark [ rmk : c ] and definition [ dfn : c ] ) . now if @xmath259 , then @xmath260 ( which we abbreviate here to @xmath261 ) is trivial , hence a coboundary . conversely , if @xmath261 is a coboundary , then there exists a @xmath2-cochain @xmath255 such that @xmath262 . thus , we have @xmath263 furthermore , @xmath264 since @xmath261 takes @xmath265 on all @xmath0-cells of @xmath266 . therefore , we have @xmath267 this completes the proof of proposition [ prop : bij ] . ( note that this last argument also shows that the map @xmath240 is indeed well - defined . ) [ rmk : act ] the proposition above implies that for each @xmath268 , the action of @xmath269 $ ] on @xmath112_0^c$ ] @xmath270_0^c\times{\mathop{\mathrm{imm}}\nolimits}[s^3 , { \mathbf{r}}^5]&\to & { \mathop{\mathrm{imm}}\nolimits}[m^3 , { \mathbf{r}}^5]_0^c\\ ( f , f)&\mapsto&f\sharp f \end{array}\ ] ] is effective and transitive . in this section , we shall define an integer invariant @xmath271_0 \to { \mathbf{z}},\ ] ] which together with the wu invariant @xmath228 will give a bijection @xmath272_0 \to \gamma_2(m^3)\times { \mathbf{z}}.\ ] ] actually we shall give two different geometric expressions for @xmath45 denoted by @xmath273 and @xmath274 which are analogues of the formulae in @xcite . then we show that they coincide . [ dfn : alpha ] let @xmath21 be a closed oriented 3-manifold . we denote by @xmath275 the dimension of the @xmath60 vector space @xmath276 , where @xmath277 is the torsion subgroup of @xmath278 . note that the number of elements in the set @xmath41 is equal to @xmath279 . the following lemma can be found in ( * ? ? ? * theorem 2.6 ) . for reader s convenience we give a short proof here . [ lem : alpha ] let @xmath21 be a closed oriented @xmath0-manifold and @xmath192 a compact oriented spin @xmath210-manifold with boundary @xmath21 . then the signature @xmath280 of @xmath192 has the same parity as @xmath281 . by @xcite , there exists a 1-connected compact oriented spin 4-manifold @xmath282 with @xmath283 such that the spin structures on @xmath21 induced by @xmath192 and @xmath282 coincide with each other . then by rohlin s theorem , @xmath284 . thus we have only to show the assertion for @xmath282 instead of @xmath192 . since @xmath282 is 1-connected , we have the exact sequence @xmath285 where the map @xmath286 is identified with the intersection form of @xmath282 through the poincar - lefschetz duality @xmath287 . by taking an appropriate basis for @xmath288 , we may assume that the matrix representative of @xmath286 is of the form @xmath289 , where @xmath290 is the zero form and @xmath291 is nonsingular . then @xmath291 can be regarded as a presentation matrix of @xmath277 and its size has the same parity as @xmath292 . now , since @xmath282 is spin , its intersection form is of even type . thus all the diagonal entries of @xmath293 are zero , where @xmath293 denotes the reduction modulo two . now , it is an easy exercise to show that the size of such a matrix has the same parity as the dimension of the @xmath60 vector space @xmath294 which it presents . thus the result follows . [ dfn : ia ] let @xmath295 be an immersion with trivial normal bundle . let @xmath296 be any compact oriented @xmath210-manifold with @xmath297 and @xmath298 a generic map nonsingular near the boundary such that @xmath299 . ( we can choose such a generic map @xmath200 , since @xmath98 is an immersion with trivial normal bundle . ) denote the algebraic number of cusps of @xmath200 by @xmath300 . then define @xmath301 we will see later that this is always an integer . [ dfn : ib ] let @xmath295 be an immersion with trivial normal bundle . let @xmath296 be any compact oriented @xmath210-manifold with @xmath297 and @xmath302 a generic map nonsingular near the boundary such that @xmath303 and @xmath304 . then define @xmath305 where @xmath306 is the algebraic number of triple points of @xmath307 , @xmath308 measures the linking of the singularity set of @xmath307 with the rest of the image @xmath309 ( see definition [ dfn : l ] ) , and @xmath310 is the linking in @xmath12 of the image @xmath108 and the double point set of @xmath98 pushed out of @xmath108 with respect to a normal framing @xmath99 for @xmath98 ( see definition [ dfn : lnu ] ) . it will be shown that @xmath274 also takes only integer values . [ lem : smootha ] the map @xmath273 is well - defined . that is , for an immersion @xmath295 with trivial normal bundle , @xmath311 does not depend on the @xmath210-manifold @xmath296 or on the choice of the generic map @xmath298 . let @xmath312 and @xmath313 be generic maps of the compact oriented @xmath210-manifolds @xmath192 and @xmath314 respectively as in definition [ dfn : ia ] . first , consider the case where the normal vector fields along @xmath108 in @xmath315 and in @xmath316 are homotopic . then after a suitable deformation near @xmath108 , we may regard @xmath317 as the image of a generic map of the closed @xmath210-manifold @xmath318 which is an immersion on a neighbourhood of @xmath21 in @xmath318 . now we obtain @xmath319 by lemma 3 in @xcite , which claims that for any generic map @xmath320 of a closed oriented @xmath210-manifold @xmath321 , the equality @xmath322 holds . thus the result follows . suppose that the normal vector fields along @xmath108 in @xmath315 and in @xmath316 are not homotopic . we may assume that both normal vector fields are of equal constant length . let @xmath323 and @xmath324 be formed by the endpoints of vectors of these two normal vector fields . then we may assume that @xmath323 and @xmath324 intersect each other along a surface @xmath325 which is the image of a surface @xmath326 by the map associating to @xmath327 the endpoint of the normal vector at @xmath328 . thus , @xmath329 can be deformed , by a small modification near @xmath108 , into the image of a smooth generic map which has only whitney umbrella singular points on @xmath330 and no other singularities nearby . now we can apply lemma 3 in @xcite again . this completes the proof . [ thm : a ] let @xmath98 and @xmath331 be two immersions with trivial normal bundles such that @xmath332 . then @xmath98 and @xmath227 are regularly homotopic if and only if @xmath333 . let us first prove that @xmath333 if @xmath98 and @xmath227 are regularly homotopic . suppose that @xmath98 and @xmath227 are regularly homotopic . let @xmath334\looparrowright{\mathbf{r}}^5\times[0,1]$ ] be the track of a generic regular homotopy @xmath335 between @xmath98 and @xmath227 . then we can take a normal vector field @xmath336 of @xmath337 , which determines normal vector fields @xmath338 and @xmath339 of @xmath340 and @xmath341 for @xmath342 and @xmath343 respectively . ( the argument below follows that of wells in @xcite . ) there is a small positive @xmath344 such that on the collars @xmath345 and @xmath346 $ ] the maps @xmath347 and @xmath348 , respectively , are immersions . let us denote these immersions by @xmath349 and @xmath350}$ ] respectively . the tangent bundle of the cylinder @xmath351 is @xmath352 , where @xmath353 $ ] , @xmath354 is the projection , and @xmath355 is the trivial line bundle . we may assume that the regular homotopy @xmath337 is such that @xmath356 for @xmath357 and @xmath358 for @xmath359 . now we define the bundle monomorphism @xmath360 as @xmath361 , i.e. for a tangent vector @xmath362 , where @xmath363 and @xmath364 , put @xmath365 . this map is monomorphic on each fibre indeed , since @xmath336 is normal to @xmath366 . furthermore , it coincides with the differentials of the immersions @xmath349 and @xmath367}$ ] over the collars . hence by the relative version of hirsch s theorem there is an immersion @xmath368 which coincides with the given immersions on the collars . let @xmath369 and @xmath370 be generic maps as in definition [ dfn : ia ] . now via the smoothing process as in the proof of lemma [ lem : smootha ] , we obtain a generic map @xmath371 ) \cup -w_g\to{\mathbf{r}}^5 $ ] of the closed oriented @xmath210-manifold @xmath372)\cup -w_g$ ] . by lemma 3 in @xcite we obtain that @xmath373 now let us show that @xmath273 is injective on the set @xmath28_0^c$ ] of regular homotopy classes of immersions having the same wu invariant @xmath203 . notice that the analogous map @xmath374 \to { \mathbf{z}}$ ] has been shown to be an isomorphism in @xcite ( see theorem [ thm2.1 ] in the present paper ) . let us consider the effective and transitive group action described in the previous section restricted to just one orbit : @xmath375_0^c \times { \mathop{\mathrm{imm}}\nolimits}[s^3,{\mathbf{r}}^5 ] \to { \mathop{\mathrm{imm}}\nolimits}[m^3,{\mathbf{r}}^5]_0^c.\ ] ] choose an embedding @xmath3 in @xmath28_0^c$ ] ( by theorem [ thm : c ] , such an embedding exists ) . let us consider the map @xmath376 & \to & { \mathop{\mathrm{imm}}\nolimits}[m^3 , { \mathbf{r}}^5]_0^c\\ f & \mapsto & f_0 \sharp f. \end{array}\ ] ] this map is bijective by proposition [ prop : bij ] . compose it with the map @xmath377_0^c \to { \mathbf{q}}.\ ] ] the resulting composition @xmath269 \to { \mathop{\mathrm{imm}}\nolimits}[m^3 , { \mathbf{r}}^5]_0^c \to { \mathbf{q}}$ ] coincides with the map @xmath378 . indeed , for any embedded seifert surface @xmath379 for @xmath3 we have @xmath380 and for a singular seifert surface of @xmath381 one can take the boundary connected sum of the seifert surface @xmath379 for @xmath3 with any singular seifert surface for @xmath59 . hence the map @xmath273 is injective . [ lem : smoothb ] for an immersion @xmath295 with trivial normal bundle and a normal framing @xmath99 , @xmath382 does not depend on the @xmath210-manifold @xmath296 or on the choice of the generic map @xmath383 . let @xmath384 and @xmath385 be generic maps of the compact oriented @xmath210-manifolds @xmath192 and @xmath314 respectively as in definition [ dfn : ib ] . then , we obtain a generic map @xmath386 . from the fact that for any generic map @xmath387 of a closed oriented @xmath210-manifold @xmath321 , the equality @xmath388 holds ( see ( * ? ? ? * lemma 4 ) ) , the result follows . the following theorem states that the invariant @xmath274 is well - defined , and it is a regular homotopy invariant . in particular it implies that @xmath382 does not depend on the choice of the normal framing @xmath99 for @xmath98 , either . [ thm : b ] let @xmath98 and @xmath331 be two immersions with trivial normal bundles such that @xmath332 . then @xmath98 and @xmath227 are regularly homotopic if and only if @xmath389 . let @xmath340 be an immersion with trivial normal bundle . by theorem [ thm : c ] , there is an embedding @xmath390 with @xmath391 . furthermore , by proposition [ prop : bij ] , there exists an immersion @xmath392 unique up to regular homotopy such that @xmath393 . let @xmath394 be the track of a generic regular homotopy between @xmath98 and @xmath381 . let @xmath395 be the algebraic number of triple points of @xmath396 . take a compact @xmath210-manifold @xmath397 and a generic map @xmath398 as in definition [ dfn : ib ] for @xmath98 , and let @xmath399 be a seifert surface for @xmath3 . then , we can deform @xmath400 into the image of a generic map of the manifold @xmath401 into @xmath402 bounded by @xmath392 ( see figure [ fig:1 ] ) . therefore , by theorem [ thm : e - s ] , we have @xmath403{0cm}{0.7cm}\frac{3}{2 } } ( \sigma ( w_f ) -\sigma ( w_{f_0 } ) + t({\widehat}{f } ) -l({\widehat}{f } ) ) + \frac{1}{2}(l(f ) - 3t(h)).\\ \end{split}\ ] ] here we need the following lemma . [ lem : l+3 t ] we have @xmath404 for any normal framing @xmath99 for @xmath98 . we can extend @xmath99 to the whole cylinder @xmath351 using the regular homotopy @xmath405 we shall denote this extended normal field also by @xmath99 . since on @xmath406 , the normal field @xmath99 coincides with the homotopically unique normal framing of @xmath407 , we have @xmath408 ( indeed , recall that for an immersion the invariant @xmath409 was defined as the linking number of the image of the immersion with the curves formed by the double points of the immersion pushed off out of the image of the immersion into the direction defined by @xmath99 . since @xmath3 is an embedding the equality follows . ) furthermore , one can show that @xmath410 in the same way as in @xcite . this completes the proof of lemma [ lem : l+3 t ] . thus , @xmath107 does not depend on the choice of @xmath99 and we obtain @xmath411{0cm}{0.7cm}\frac{1}{2}}(3\sigma ( w_f ) + 3t({\widehat}{f } ) -3l({\widehat}{f } ) + l_\nu(f ) ) - \frac{3}{2}\sigma ( w_{f_0 } ) . \end{split}\ ] ] since the regular homotopy class of @xmath412 is determined by @xmath157 , again by proposition [ prop : bij ] , the regular homotopy class of @xmath98 is completely determined by @xmath413 , hence by @xmath414 . this completes the proof of theorem [ thm : b ] . [ rmk : lindep ] theorem [ thm : b ] implies that @xmath382 does not depend on the choice of @xmath307 or on the normal framing @xmath99 for @xmath98 . hence @xmath310 does not depend on the choice of the normal framing @xmath99 for @xmath98 , either . thus , we can use the notation @xmath415 for an immersion @xmath98 with trivial normal bundle of a closed oriented @xmath0-manifold from now on ; then , the formula displayed in definition [ dfn : ib ] can be written as @xmath416 [ rmk : coin ] we see easily that for an embedding @xmath417 we have @xmath418 ; furthermore , this is always an integer by lemma [ lem : alpha ] . therefore , by theorems [ thm : a ] , [ thm : b ] , [ thm2.1 ] and [ thm : e - s ] , @xmath273 and @xmath274 take integer values for general immersions ( with trivial normal bundles ) . furthermore , considering the action of @xmath26 $ ] on @xmath28_0 $ ] , we see that in fact @xmath419 ( using theorems [ thm2.1 ] and [ thm : e - s ] ) . thus , we obtain the surjective map @xmath420_0\longrightarrow{\mathbf{z}}\ ] ] defined in two ways as @xmath273 and as @xmath274 , so that the map @xmath421_0\longrightarrow\gamma_2(m^3 ) \times{\mathbf{z}}\ ] ] gives a bijection . [ cor : emb ] let @xmath21 be a closed oriented @xmath0-manifold such that @xmath23 has no @xmath2-torsion . then two embeddings @xmath22 are regularly homotopic if and only if they have seifert surfaces with the same signature since @xmath23 has no @xmath2-torsion , @xmath157 is equal to @xmath422 for every embedding @xmath423 . if we take a seifert surface @xmath424 for an embedding @xmath98 , then clearly @xmath425 . thus the result follows directly from theorems [ thm : a ] and [ thm : b ] . as a consequence of theorems [ thm : a ] and [ thm : b ] , we have that _ if two embeddings @xmath98 and @xmath426 of an arbitrary closed oriented @xmath0-manifold @xmath21 into @xmath12 are regularly homotopic , then their seifert surfaces have the same signature_. in fact , we can also prove this fact by using a simpler argument as follows . let @xmath427 and @xmath428 be seifert surfaces for @xmath98 and @xmath227 respectively . let @xmath429 be the track of a generic regular homotopy between @xmath98 and @xmath227 . then , using @xmath200 , @xmath430 , and @xmath396 , we obtain a generic immersion @xmath431 after an appropriate smoothing process . put @xmath432 . let @xmath433 be the normal euler class of the immersion @xmath11 and @xmath434 its poincar dual . then , @xmath435 and @xmath436 is represented by the set of double points @xmath437 of @xmath11 ( see ( * ? ? ? * , lemmas 1 and 4 ) ) . since @xmath438 , we have [ cols="^,^ , < " , ] + thus , we have @xmath439 and hence @xmath440 . in this section , we prove the two surprising corollaries formulated in the introduction . recall that these corollaries claim roughly that for the @xmath0-dimensional torus @xmath5 `` more regular homotopy classes of immersions into @xmath12 contain embeddings than for the @xmath0-sphere '' . ( the proofs will show that there are twice as many regular homotopy classes containing embeddings for the torus as for the sphere . ) since @xmath441 , the set @xmath442_0 $ ] can be identified with the set of integers . therefore , by proposition [ prop : bij ] , we see that @xmath443\longrightarrow { \mathop{\mathrm{imm}}\nolimits}[t^3 , { \mathbf{r}}^5]_0\ ] ] is a bijection for any immersion @xmath444 with trivial normal bundle . furthermore , @xmath445_0\to{\mathbf{z}}$ ] in section [ sect : main ] gives a complete invariant of regular homotopy . obviously there exists an embedding @xmath446 which bounds a @xmath210-manifold of signature @xmath265 ( e.g. @xmath447 ) in @xmath12 . furthermore , we show the following . [ prop : t3 ] * there exists an embedding @xmath448 having a seifert surface @xmath449 with signature @xmath450 . * the signature of any seifert surface of any embedding @xmath451 is divisible by @xmath450 . \(a ) it is known that there exists a spin @xmath210-manifold @xmath452 of signature @xmath450 with @xmath453 , see @xcite for example . furthermore , @xmath192 can be chosen to have a special handlebody decomposition with one @xmath265-handle and some @xmath2-handles with even framings ( see @xcite ) . as we have seen in the proof of theorem [ thm : c ] , such a @xmath210-manifold @xmath192 embeds into @xmath12 . such an embedding , restricted to the boundary @xmath454 , gives a required embedding @xmath455 . \(b ) it is shown also in @xcite that @xmath5 with any spin structure spin - bounds either the solid torus @xmath456 or the above mentioned @xmath457 . now let @xmath282 be any seifert surface of any embedding @xmath451 . then @xmath282 induces a spin structure on @xmath5 , which is induced also from a spin structure either on the solid torus or on @xmath457 . therefore , either @xmath458 or @xmath459 carries a spin structure , and so by rohlin s theorem its signature is divisible by @xmath460 . by novikov s additivity these signatures are @xmath461 and @xmath462 respectively . in both cases @xmath292 is divisible by @xmath450 . [ cor : t3sz ] there exist two embeddings @xmath3 and @xmath463 such that there is an immersion @xmath464 with @xmath465 , but @xmath11 can not be chosen from a regular homotopy class containing an embedding . let @xmath3 and @xmath8 be the embeddings described above in proposition [ prop : t3 ] ; i.e. they have seifert surfaces@xmath379 and @xmath457 of signatures @xmath466 and @xmath467 respectively . from the fact that@xmath443\longrightarrow { \mathop{\mathrm{imm}}\nolimits}[t^3 , { \mathbf{r}}^5]_0\ ] ] is a bijection , there is an immersion @xmath468 such that @xmath465 . let us assume that @xmath11 is regularly homotopic to an embedding . then the following sequence of equalities hold : @xmath469 for an integer @xmath470 , which is a contradiction . here the fourth equality @xmath471 follows from the obvious remark that for a seifert surface of the connected sum @xmath7 one can choose the boundary connected sum of the seifert surfaces of the embeddings @xmath3 and @xmath11 . the smale invariant @xmath472 is a multiple of @xmath473 by the result of hughes and melvin @xcite recalled in section [ sect : rev ] . [ cor : t3 ] there exists an immersion @xmath9 not regularly homotopic to any embedding @xmath474 such that for any embedding @xmath475 , the connected sum @xmath476 of @xmath98 and @xmath11 is again regularly homotopic to an embedding @xmath451 . let @xmath9 be any immersion with smale invariant @xmath477 . then @xmath11 is not regularly homotopic to an embedding again by the result of hughes - melvin @xcite . we show that for any embedding @xmath478 the connected sum @xmath476 is regularly homotopic to an embedding . for this purpose we first compute the @xmath45-invariant of @xmath476 and then produce an embedding with the same @xmath45-invariant . since the @xmath45-invariant determines the regular homotopy class completely , we obtain that @xmath476 is regularly homtopic to the embedding . the following sequence of equalities hold : @xmath479 here the first equlaity holds , since we can form the boundary connected sum of the seifert surface @xmath480 of @xmath481 with that of @xmath11 even if the latter is a `` singular '' one ( i.e. it is a generic map with boundary @xmath11 , not necessarily an embedding ) . the number @xmath482 is an integer by part ( b ) of proposition [ prop : t3 ] . let @xmath483 be an embedding with smale invariant @xmath484 , where @xmath485 if @xmath486 is even and @xmath487 if @xmath486 is odd . now we see that if @xmath486 is even , then @xmath488 . if @xmath486 is odd , then @xmath489 . the present section is not along the main line of the paper . here as a byproduct of our formulae for @xmath273 and @xmath274 , their correctness , integrality and coincidence , we prove a few simple corollaries on the number of cusps a singular seifert surface can have . [ cor : cusp1 ] in the notation of definition [ dfn : ia ] , the residue class of the algebraic number @xmath490 of cusps of @xmath491 modulo @xmath0 does not depend on the @xmath210-manifold @xmath296 or on the choice of the map @xmath491 , and it depends only on the immersion @xmath98 of the @xmath0-manifold @xmath21 into @xmath12 . naturally arises the problem : how to read off this modulo @xmath0 residue class from @xmath98 ? it turns out that @xmath492 ( see definition [ dfn : lnu ] ) . this follows from the definitions of the invariants @xmath273 and @xmath274 ( definitions [ dfn : ia ] and [ dfn : ib ] ) and their coincidence ( see remark [ rmk : coin ] ) . [ cor : cusp2 ] let @xmath296 be a closed connected spin @xmath210-manifold , and @xmath493 a generic map . it is well - known that the singularity set is a surface @xmath494 with isolated cusps on it . then the number of cusps on each component of @xmath494 is even . otherwise , restricting @xmath59 to the closure of an appropriate neighbourhood of a component of @xmath494 , we would get a contradiction to the integrality of the invariant @xmath273 and lemma [ lem : alpha ] . [ cor : cusp3 ] in the conditions of the previous corollary , let @xmath21 be a null - homologous oriented submanifold of @xmath296 disjoint from @xmath495the closure of@xmath496 the double point set of @xmath59 . then in both parts of the set @xmath497 the algebraic number of cusps is divisible by @xmath498 . 99 t. ekholm , _ regular homotopy and vassiliev invariants of generic immersions @xmath499 , @xmath500 _ , j. knot theory ramifications * 7 * ( 1998 ) , 10411064 . t. ekholm , _ differential @xmath0-knots in @xmath1-space with and without self intersections _ , topology * 40 * ( 2001 ) , 157196 . t. ekholm , _ geometry of self intersection surfaces and vassiliev invariants of generic immersions @xmath501 _ , to appear in algebra and analysis . t. ekholm and a. szcs , _ geometric formulas for smale invariants of codimension two immersions _ , preprint ( arxiv : math.gt/0002076 ) . m. hirsch , _ immersions of manifolds _ , trans . amer . math . * 93 * ( 1959 ) , 242276 . j. hughes , _ bordism and regular homotopy of low - dimensional immersions _ , pacific j. math . * 156 * ( 1992 ) , 155184 . j. hughes and p. melvin , _ the smale invariant of a knot _ , comment . * 60 * ( 1985 ) , 615627 . s. j. kaplan , _ constructing framed @xmath210-manifolds with given almost framed boundaries _ , * 254 * ( 1979 ) , 237263 . r. kirby , _ the topology of @xmath210-manifolds _ , lecture notes in math . * 1374 * ( springer - verlag , 1989 ) . r. kirby and p. melvin , _ canonical framings for @xmath0-manifolds _ , proceedings of 6th gkova geometry - topology conference , turkish j. math . * 23 * ( 1999 ) , 89115 . bang - he li , _ on classification of immersions of @xmath13-manifolds in @xmath502-manifolds _ , comment . * 57 * ( 1982 ) , 135142 . s. smale , _ a classification of immersions of spheres in euclidean spaces _ , ann . of math . * 69 * ( 1959 ) , 327344 . a. szcs , _ on the singularities of hyperplane projections of immersions _ , bull . london math . * 32 * ( 2000 ) , 364374 . m. takase , _ embeddings of @xmath60-homology @xmath0-spheres in @xmath12 up to regular homotopy _ , pacific j. math . * 193 * ( 2000 ) , 249256 . r. wells , _ cobordism of immersions _ , topology * 5 * ( 1966 ) , 281294 . h. whitney , _ the self - intersections of a smooth @xmath13-manifold in @xmath503-space _ , ann . of math . * 45 * ( 1944 ) , 220246 . h. whitney , _ classification of the mappings of a @xmath0-complex into a simply connected space _ , ann . of math . * 50 * ( 1949 ) , 270284 . wen - tsn wu , _ on the immersion of @xmath504-@xmath0-manifolds in a euclidean space _ , sci . sinica * 13 * ( 1964 ) , 335336 .	we give geometric formulae which enable us to detect ( completely in some cases ) the regular homotopy class of an immersion with trivial normal bundle of a closed oriented @xmath0-manifold into @xmath1-space . these are analogues of the geometric formulae for the smale invariants due to ekholm and the second author . as a corollary , we show that two embeddings into @xmath1-space of a closed oriented @xmath0-manifold with no @xmath2-torsion in the second cohomology are regularly homotopic if and only if they have seifert surfaces with the same signature . we also show that there exist two embeddings @xmath3 and @xmath4 of the @xmath0-torus @xmath5 with the following properties : @xmath6 @xmath7 is regularly homotopic to @xmath8 for some immersion @xmath9 , and @xmath10 the immersion @xmath11 as above can not be chosen from a regular homotopy class containing an embedding .
the corrosion of metalic surfaces is a subject of increasing interest due to widespread use of corrodable materials with all the economical and environmental implications @xcite . when the metal is in contact with an aggressive solution , the formation of a passive layer containing the metal oxide is observed after the corrosion of the topmost layer of the surface . as this layer grows , it more efficiently protects the surface from the attack of the solution , leading to the slowdown of the corrosion rate . for instance , according to the cabrera and mott@xcite law , the corrosion front is expected to move as @xmath9 ( growth rate decreasing as @xmath2 ) , which is in agreement with experimental findings . the changes in the corrosion rate are certainly accompanied by changes in the roughness of the front separating the metal and the passive layer . in the beginning of the process , all the surface is uniformly exposed to the corrosive solution , but after the formation of the passive layer there appear points of easier or more difficult access to the solution . it is particularly important to investigate the scaling properties of the roughness because they are intimately connected to the main dynamical processes during the interface evolution @xcite . from the theoretical point of view , additional motivation comes from the renewed interest in systems whose competitive dynamics leads to crossover in roughness scaling @xcite . here we will analyze the crossover in the front evolution and in the roughness scaling of a recently introduced model for the corrosion of a metalic surface by a solution @xcite . the basic mechanisms of the model are the corrosion of the mettalic surface exposed to the solution and the formation of the passive layer with the oxides . the layer is assumed to be porous , but the larger molar volumes of the products may block the pores and , consequently , diffusive relaxation of the volume excess is necessary for new corrosion events . this model was not devised to represent a particular real process @xcite , but to contain essential ingredients of a large variety of corrosion processes , allowing the investigation of universal features . previous numerical work on this model has shown significant differences in the corrosion fronts for small and large values of the local corrosion rate of the metal in contact with the solution @xcite . however , besides these qualitative features , only universal laws for the global growth rates in limiting cases were already derived @xcite . in the present paper , we advance the quantitative study of the model , focusing on the properties of the corrosion front which separates the metal and the passive layer . we show the emergence of a crossover in the global growth rate accompanied by a crossover in surface roughness scaling , from a short time kpz behavior to long time laplacian growth ( or diffusion - limited erosion - dle ) @xcite . a scaling picture of the process is derived , along the same lines of those developed for related systems @xcite , and the results are confirmed by numerical simulation . the rest of this paper is organized as follows . in sec . ii we present the model for corrosion and passivation . in sec . iii , we present a scaling theory for the front growth and numerical results that confirm it . in sec . iv , we discuss the crossover kpz - dle of the surface roughness . in sec . v we summarize our results and present our conclusions . the model represents a system with the structure shown in fig . while the metal is corroded , a porous passive layer grows , with some products filling their pores . these products block the access of the solution to the surface , but they may diffuse and leave the surface of the metal exposed again . for simplicity , the model is defined in a lattice , where the sites may have three different states : m ( metal ) , s ( solution ) or p ( passive ) . the lattice sites represent the dominant chemical species inside one a mesoscopic region . due to computational limitations of memory and time , a square lattice will be considered in our numerical work . in order to represent the effect of the porosity of the passive layer , we assume that a site m can be corroded ( passivated ) even when it is in contact only with p sites , thus avoiding the complication of representing the porous structure of the passive layer . the excess of products will be represented by particles w , which can only lay at p sites and may diffuse through the p matrix . at most one particle w is allowed at each p site . a region of the lattice is illustrated in fig . 1b . in the beginning of the process , all sites above a certain row of the lattice ( height @xmath10 ) are labeled s , and all sites below or at that row are labeled m. an m site is said to be in contact with the solution if it has at least one neighbor s or one neighbor p without a w particle over it . at each time unit , each m satisfying this condition may be changed into p with probability @xmath0 ( reaction @xmath11 ) . it corresponds to the reduction of a chemical species in the solution by the metal and the formation of an insoluble film . a reaction @xmath11 is illustrated for the central site in fig . 2a . on the other hand , 2b shows a case where this reaction is not possible for the central m site due to the blocking effect of the w particle . we assume that the molar volume of the oxide products is larger than that of the metal ( pilling and bedworth factor larger than one @xcite ) , and we represent this excess by creation of w particles after the reaction @xmath11 . two w particles are produced immediately after each reaction : the first one lays at the newly created p site and the second one lays at a neighboring p or s site . the creation of the new w particles is also illustrated in fig . some differences between models with creation of one or two w particles per reaction were discussed in refs . @xcite . the particles w may also diffuse through the p layer . this represents some type of stress relaxation which leads to the redistribution of the excess of corrosion products @xcite . at each time unit , each w attempts to execute @xmath1 random displacements to neighboring sites . the movement is possible only if the target site is p or s and if it does not contain a w particle , otherwise the attempt is rejected . an acceptable movement is illustrated in fig . 2c , but an attempt to move the same w particle to any other direction would be rejected in that case . when a w particle attempts to move to an s site , or when the w particle is created over an s site ( top of the passive layer ) , this s site is converted into p. this condition is reasonable to represent the expansion of the passive layer towards the solution side . however , it does not have a significant influence on the corrosion front , which separates the metal and the passive layer , and which is the main interest of the present study . here we focus on the case of small corrosion rates @xmath0 and of not very small values of @xmath1 . first we consider the short time behavior of the system . in one time unit , the number of corroded sites is small compared to the total number of sites in the front , but all w particles attempt to move during the same time interval . the diffusion of w particles removes the restrictions for the corrosion and the situation in fig . 2b is rare : the w particle at the neighboring p site will move away long before a successful corrosion attempt occurs at the central site . consequently , we expect that all m sites at the interface are available for corrosion with equal probabilities . the distance of the corrosion front from the initial level , hereafter called the height of the corrosion front , increases as @xmath12 ( @xmath13 is being measured in lattice units ) . in this regime , the model is equivalent to an inverse of the eden model @xcite , in which a new particle is added to a randomly chosen neighbor of the aggregate surface at each time step . now consider the system at long times . since w particles move diffusively , their upwards displacement away from the corrosion front increases as @xmath14 , thus their velocity in the upwards direction evolves as @xmath15 \sim { \left ( d / t\right)}^{1/2}$ ] . this decreasing velocity leads to the accumulation of w particles near the surface , which block new corrosion events . this is schematically illustrated in fig.3 , where a sea of w particles is shown at the bottom of the passive layer . the crossover from the initial regime to the long time one takes place at a characteristic time @xmath8 in which the height of the sea of w particles covers a finite fraction of the passive layer . the height of that sea increases as @xmath16 because it is determined by the diffusion of w particles produced at short times . on the other hand , the height of the passive layer increases as eq . ( [ hini ] ) before the crossover . matching these heights we obtain the crossover time @xmath17 the height of the corrosion front at the crossover , measured in lattice units , is @xmath18 after the crossover , a corrosion event is possible only when an m site has an empty neighbor p , which requires the migration of voids ( holes ) through the sea of w particles and their arrival at the corrosion front . when the m site is corroded ( reaction @xmath11 ) , the hole of the w sea disappears because it is occupied by one of the new w particles . since these holes move diffusively from the top of the w sea to that front ( similarly to the w particles ) , we conclude that that the corrosion front will also move diffusively , with displacement increasing as @xmath9 and velocity decreasing as @xmath2 . the time evolution of the height @xmath19 of the corrosion front is expected to follow a scaling relation that involves @xmath20 and @xmath8 : @xmath21 where @xmath22 is a scaling function . indeed , with @xmath23 for small times , we regain eq . ( [ hini ] ) , and with @xmath24 for long times we obtain @xmath25 , in agreement with the above discussion . in order to test the scaling relation ( [ scalingh ] ) , we use simulation data for several values of @xmath0 , ranging from @xmath26 to @xmath27 , and @xmath28 . these data were obtained in lattices with lateral length @xmath29 and periodic boundary conditions . in fig . 4 we show a scaling plot of @xmath30 versus @xmath31 , which are respectively proportional to @xmath5 and @xmath6 . the excellent data collapse of the results for consecutive values of @xmath0 shows that there is a universal relation between those quantities . lines with slopes @xmath32 and @xmath33 are also included in the plot in order to highlight the crossover from the @xmath34 growth to the @xmath35 growth . in previous numerical works with small and large @xmath0 @xcite , this universal behavior was not observed because , for a single value of @xmath0 , it is difficult to span a time interval long enough to show both scaling regimes . numerical studies in three dimensions would certainly be more difficult . however , the scaling picture will be the same in three dimensions because it is derived from diffusion properties and the same stochastic rules for the model . for small times , the equivalence to the eden model ( sec . iii ) indicates that the corrosion model is in the kpz class of growth @xcite , i. e. in the continuous limit it is represented by the kpz equation @xmath36 here , @xmath19 is the height at the position @xmath37 in a @xmath38-dimensional substrate at time @xmath39 , @xmath40 and @xmath41 are constants and @xmath42 is a gaussian noise @xcite . in the growth regime , the lateral correlation length is much smaller than the lattice length and the roughness increases as @xmath43 , where @xmath44 is called the growth exponent ( @xmath45 in @xmath46 @xcite , @xmath47 in @xmath48 @xcite ) and @xmath49 is an amplitude related to the coefficients of the kpz equation . in the original eden model , each site of the front can grow with probability @xmath32 at each time unit @xcite . however , the corrosion of an m site in our model takes place with probability @xmath0 per time unit . thus , for small times we expect that @xmath50 where @xmath51 is a constant related to eden growth properties . in fig . 5 we show a log - log plot of the small time roughness as a function of @xmath52 , for small @xmath0 , confirming the validity of eq . ( [ wkpz ] ) . the average slope of this plot is significantly below the asymptotic kpz value @xmath45 , but these short time deviations are expected in the eden model @xcite as well as in related kpz models @xcite . at long times , a corrosion event takes place when a void of the w sea arrives at the metal surface , after performing a random walk which begins at a distant point . consequently , surface peaks will be more easily subject to corrosion than the valleys of the surface , which favors the production of surfaces with low roughness , as observed in previous numerical works for high values of @xmath0 @xcite . these features indicate that the long time regime is equivalent to that of the dle model @xcite . in dle , eroding particles are randomly left at points very distant from the solid surface , they are allowed to diffuse and , when they reach the surface , they are annihilated together with the particle of the solid . in our model , the eroding particles of dle are represented by the holes of the w sea . in the continuous description of dle , the stable interface is driven by the gradient of a laplacian field @xcite . a logarithmic scaling of the squared roughness is predicted in @xmath53 dimensions : @xmath54 where @xmath55 is the growth velocity , @xmath39 is the time , @xmath56 is the lattice parameter and @xmath57 , where @xmath58 is a lengthscale associated with the noise @xcite . our model is equivalent to a dle with a probability @xmath0 of the incident particle to remove a particle of the solid . moreover , changes in the time scale are also expected because the growth rate of our model decreases in time , while it is constant in the original dle @xcite . thus , for a connection of the dle roughness in eq . ( [ wlaplacian ] ) and that of our model , we use @xmath59 as the height of the corrosion front and @xmath60 . this gives @xmath61 where @xmath19 is measured in lattice units and @xmath62 is a time - independent constant associated with the noise ( @xmath0 and @xmath1 ) . the time scaling of the surface roughness , implicitly given in its relation with the height @xmath19 in eq . ( [ wdle ] ) , is tested in fig . 6 , where we plot @xmath63 as a function of @xmath31 . large values of @xmath31 were chosen in order to ensure that only data in the long time regime were considered ( @xmath64 - see fig . 4 ) , but simulations for extremely long times are feasible only for large @xmath0 . we notice that , for a given value of @xmath0 , @xmath65 oscillates around some constant value , which provides the estimate of the constant @xmath62 in eq . [ wdle ] . the crossover from kpz to dle scaling is expected to take place at a time of order @xmath8 because the changes in the growth rate and in the roughness scaling are caused by the same mechanisms . in other models with crossover between different growth dynamics , it is observed that matching the scaling relations for the roughness of both dynamics leads to correct predictions of @xmath8 @xcite . following this reasoning , the @xmath66 values in eqs . ( [ wkpz ] ) and ( [ wdle ] ) must be the same at time @xmath67 and for a height @xmath68 . this gives @xmath69 where @xmath70 and @xmath71 are constants . in order to test relation ( [ eqb ] ) , we plot @xmath72 as a function of @xmath73 in fig . we obtain a reasonable linear fit of the data for three orders of magnitude of @xmath0 , which suggests the validity of our approach . the crossover of surface roughness is slightly different in @xmath74 dimensions because the exponent @xmath44 of the kpz class is changed @xcite and the roughness of dle saturates at a finite value @xcite . however , the model rules indicate that the short and long time dynamics are also kpz and laplacian growth , respectively . we studied the crossover from kardar - parisi - zhang growth to laplacian front growth in a model for corrosion and passivation of a metalic surface . at small times , a regime of constant growth of the corrosion front is observed , with the roughness obeying kpz scaling . at long times , a large fraction of the layer is blocked , leading to a global corrosion rate that decreases as @xmath2 . the roughness shows logarithmic scaling , with amplitudes depending on the corrosion probability @xmath0 . scaling arguments show that the crossover between these different dynamics takes place at @xmath67 , and that the height of the corrosion front at that time is @xmath68 . universal scaling relations between the height of the corrosion front @xmath19 and the time @xmath39 are obtained , as well as relations for the surface roughness . all results are confirmed by numerical simulations in square lattices , although the scaling arguments are also valid in three - dimensions . as far as we know , the crossover from kpz to dle scaling in statistical growth models was not predicted before . the inverse situation , i. e. the crossover from dle ( short times ) to kpz ( long times ) was suggested to appear in an extension of the dle model where the movement of the eroding particles is biased towards the surface @xcite . however , recent simulation results suggest a crossover to edwards - wilkinson scaling instead of kpz @xcite . we also recall that the crossover of our model must not be confused with the widely studied crossover from kpz to diffusion - limited aggregation ( dla ) - see e. g. ref . @xcite . indeed , surface roughness scaling is very different in dle and dla . the universal behavior discussed here may be observed in real systems where the basic features illustrated in fig . 1 are expected . such tests are certainly possible with the modern techniques of microscopy , that allow the determination of surface roughness with good accuracy . one of the interesting points for comparison is the time behavior of the roughness , which evolves from a rapid growth , typical of kpz scaling , to a very slow growth , typical of the logarithmic scaling in laplacian front growth . moreover , it would also be interesting to confirm the association of the changes in the growth rate with the changes in roughness scaling .	we study a model of corrosion and passivation of a metalic surface in contact with a solution using scaling arguments and simulation . the passive layer is porous so that the metal surface is in contact with the solution . the volume excess of the products may suppress the access of the solution to the metal surface , but it is then restored by a diffusion mechanism . a metalic site in contact with the solution or with the porous layer can be passivated with rate @xmath0 and volume excess diffuses with rate @xmath1 . at small times , the corrosion front linearly grows in time , but the growth velocity shows a @xmath2 decrease after a crossover time of order @xmath3 , where the average front height is of order @xmath4 . a universal scaling relation between @xmath5 and @xmath6 is proposed and confirmed by simulation for @xmath7 in square lattices . the roughness of the corrosion front shows a crossover from kardar - parisi - zhang scaling to laplacian growth ( diffusion - limited erosion - dle ) at @xmath8 . the amplitudes of roughness scaling are obtained by the same kind of arguments as previously applied to the other competitive growth models . the simulation results confirm their validity . since the proposed model captures the essential ingredients of different corrosion processes , we also expect these universal features to appear in real systems .
since the lifshitz - kosevich theory @xcite provided a basis for universal quantitative description of magnetic quantum oscillations , these effects became one of most popular experimental means of studying the fermi surface properties of metals @xcite . besides traditional metals , quantum oscillations of magnetoresitance ( shubnikov - de haas , sdh effect ) and magnetization ( de haas - van alphen effect ) have recently proved extremely useful in exploring more complex topical materials such as cuprate @xcite and iron - based superconductors @xcite , topological insulators @xcite , heavy fermion compounds @xcite , and organic charge - transfer salts @xcite . here we report on an experimental study of the high - field interlayer magnetoresistance of the layered conductor @xmath0-(bets)@xmath1mn[n(cn)@xmath1]@xmath3 , demonstrating the power of the sdh effect in exploring fermi surface properties of a quasi - two - dimensional correlated electronic system . the present compound belongs to the family of bifunctional organic charge - transfer salts , in which conducting and magnetic properties are formed by different electronic subsystems spatially separated on a subnanometer level . the electrical conductivity is provided by delocalized @xmath4 electrons of fractionally charged bets donors arranged in two - dimensional ( 2d ) sheets , whereas magnetic properties are dominated by localized @xmath5-electron spins of mn@xmath6 in the insulating anionic layers @xcite . in addition to the interesting , still not understood crosstalk between the two subsystems @xcite , the narrow , half - filled conduction band is a likely candidate for a mott instability @xcite . the material undergoes a metal - insulator transition at @xmath7k @xcite . the insulating ground state is very sensitive to pressure : under a quasi - hydrostatic pressure of about 1kbar it is completely suppresses , giving way to a metallic and even a superconducting state with @xmath8k @xcite . a thorough knowledge of the fermi surface and other basic properties of the normal - state charge carriers is certainly mandatory for understanding the interplay between the various instabilities of the normal metallic state . to this end we have carried out high - field magnetoresistance studies of @xmath0-(bets)@xmath1mn[n(cn)@xmath1]@xmath3 under pressure @xmath9kbar . this pressure drives the system in the metallic part of the phase diagram , however , not far away from the insulating domain . we have found sdh oscillations with two fundamental frequencies , indicating a fermi surface consistent with the predictions of the band structure calculations @xcite . our analysis of the oscillation behaviour suggests strong electron correlations which are considerably dependent on the inplane wave vector . single crystals of @xmath0-(bets)@xmath1mn[n(cn)@xmath1]@xmath3 were grown electrochemically , as described elsewhere @xcite and had a shape of small plates with characteristic dimensions of @xmath10mm@xmath11 . the largest surface of the plate was parallel to the highly conducting bets molecular layers , which is defined as the crystallographic @xmath12-plane . resistive measurements were done with a standard four - probe a.c . technique using a low - frequency ( @xmath13hz ) lock - in amplifier . two contacts were attached to each of two opposite sample surfaces with conducting graphite paste in order to measure the interlayer resistance . the magnetoresistance measurements were done in the temperature range @xmath14k in magnetic fields of up to 29 t generated by the 24mw resistive magnet at the lncmi - grenoble . a quasi - hydrostatic pressure of 1.4kbar was applied using the cu - be clamp cell with silicon oil as a pressure medium and with a manganin coil for pressure control . the samples were aligned with the normal to conducting layers ( crystallographic @xmath15-axis ) being parallel to the magnetic field . -(bets)@xmath1mn[n(cn)@xmath1]@xmath3 crystal as a function of magnetic field perpendicular to the layers , at @xmath16k . inset : close - up view of the oscillatory component of the field - dependent resistance normalized to the monotonic background : @xmath17 . the background @xmath18 was determined by a low - order polynomial fit to the as - measured resistance @xmath19.,scaledwidth=45.0% ] figure [ sdh ] shows the low - temperature interlayer resistance of a pressurized @xmath0-(bets)@xmath1mn@xmath3 crystal measured in a magnetic field perpendicular to the layers . on the background of a monotonic , almost linear magnetoresistance one can see prominent sdh oscillations . the fast fourier transform ( fft ) of the oscillatory signal is presented in fig.[fft ] for two field windows , 12 to 15 t and 23 to 29 t . in both spectra the dominant frequency is @xmath20 t . in addition , a smaller peak is observed at @xmath21 t and , in the higher - field spectrum , at @xmath22 t , which is the second harmonic of @xmath23 . the contribution from the @xmath24 frequency is stronger at relatively low fields , @xmath25 t and can be seen by bare eye , for example , in the inset in fig.[sdh ] . at increasing the field its relative contribution decreases ; however , it is still present in the high - field fft spectrum in fig.[fft ] . k taken in the field windows @xmath26 t ( lower curve ) and @xmath27 t ( upper curve ) . two peaks at the fundamental frequencies @xmath28 and @xmath23 correspond , respectively , to the classical ( @xmath24 ) and mb ( @xmath29 ) orbits on the 2d fermi surface @xcite shown in the upper left corner.,scaledwidth=45.0% ] the @xmath29 frequency reveals a cyclotron orbit area in @xmath30-space equal to the first brillouin zone ( bz ) area . this result is in full agreement with the band structure calculations @xcite , predicting a 2d fermi surface as shown in the upper left inset in fig.[fft ] . the corresponding cyclotron orbit in the extended zone scheme is shown in the upper right corner in fig.[fft ] . due to the inversion symmetry of the molecular layer , the band dispersion was predicted to be degenerate at the fermi level on the line z m of the bz boundary @xcite . in that case only one fundamental sdh frequency @xmath23 would be expected . however , we clearly observe the frequency @xmath28 corresponding to an orbit occupying @xmath31 of the bz area . this is exactly the size of the @xmath24 orbit centered at point z of the bz boundary , see fig.[fft ] . therefore , we conclude that the band degeneracy is lifted , most likely due to a finite spin - orbit interaction @xcite . the resulting fermi surface should consist of a closed part @xmath24 and a pair of open sheets extended along @xmath32 ( parallel to @xmath33 z line in fig.[fft ] ) . the large orbit @xmath29 is then a consequence of magnetic breakdown ( mb ) through the small gap between the different parts of the fermi surface . similar evidence of the mb effect were found on several well known @xmath0-(bedt - ttf)@xmath1x salts @xcite . the sdh spectrum in fig.[fft ] differs from that reported earlier for this compound at similar pressures @xcite . in the earlier experiment no @xmath28 and @xmath23 but , instead , very slow oscillations with the frequency @xmath34 t were found . while the absence of the @xmath24 and @xmath29 oscillations can easily be attributed to the higher temperatures and lower fields used in ref . , the absence of the slow oscillations in the present experiment is most likely caused by a different pressurizing procedure . in the previous work the samples were cooled at ambient pressure and then a pressure of @xmath35kbar was applied at temperatures below 20k , using the he - gas pressure technique . during the ambient pressure cooling , a superstructure transition was detected at @xmath36k @xcite . this transition was proposed to give rise to small fermi pockets responsible for the slow oscillations . by contrast , in the present work we apply pressure already at room temperature , using the clamp cell technique . the pressure probably prevents the formation of superstructure and the associated reconstruction of the fermi surface . indeed , the characteristic transition feature found in the ambient - pressure @xmath37 cooling curves has not been detected under pressure . we now turn to quantitative analysis of the oscillations . generally speaking , the quasi-2d character of the electronic system may lead to strong violations of the standard lifshitz - kosevich ( lk ) theory , see , e.g. , refs .. however , if we compare the present oscillations with those observed on some other highly 2d , clean organic metals @xcite , their amplitude is relatively weak and the harmonic content is low . as will be shown below , the oscillations exhibit the conventional exponential temperature and field dependence . therefore , in the following we apply the conventional lk formalism described in detail in ref .. this approach has been proved to give reasonable results for other similarly anisotropic organic metals at not too high magnetic fields @xcite . we consider the relative amplitudes of the oscillations in the form @xcite : @xmath38 where the subscript @xmath39 labels the relevant orbit on the fermi surface , @xmath40 is the amplitude of the oscillations in conductivity normalized to the respective nonoscillating background , @xmath41 is the field- and temperature - independent prefactor and @xmath42 , @xmath43 , and @xmath44 are the damping factors caused , respectively , by finite temperature , scattering , and mb effects . the temperature dependence of the sdh amplitude can be fitted by the lk temperature damping factor @xcite : @xmath45 where @xmath46t / k and @xmath47 is the effective cyclotron mass in units of the free electron mass @xmath48 . for a large argument of the @xmath49-function in eq.([rt ] ) the logarithmic plot of the ratio @xmath50 , known as the lk plot , should be a straight line with a slope proportional to @xmath51 . this is , indeed , true for both the @xmath24 and @xmath29 oscillations , as demonstrated in fig.[mass ] . fitting the slopes yields the cyclotron masses @xmath52 and @xmath53 . ( circles ) and @xmath54 ( squares ) . the amplitudes were determined by fft in the field window 14 to 17 t . the lines are fits by eq.([rt ] ) with normalized cyclotron mass values of 5.53 and 6.88 for the @xmath24 and @xmath29 oscillations , respectively.,scaledwidth=45.0% ] the obtained values can be compared to the band masses estimated within a noninteracting electron model . to this end , we have carried out tight - binding band structure calculations based on the organic donor homos ( highest occupied molecular orbitals ) obtained by the extended hckel method @xcite . to estimate the energy dependence of the fermi surface area and thus the cyclotron mass , the calculations were done for different band fillings near the fermi level with a step of @xmath55 electron per bets donor . the resulting values are @xmath56 and @xmath57 . although these estimations are rather rough , they demonstrate that the real cyclotron masses are considerably enhanced , most likely due the many - body effects . the significance of electronic correlations would not be very surprising , since @xmath0-(bets)@xmath1mn[n(cn)@xmath1]@xmath3 is proposed to be a mott insulator at ambient pressure @xcite and at the given pressure , @xmath9 kbar , the compound is still quite close to the metal - insulator boundary . it should be noted that the present mass values , especially @xmath58 , are even higher than those reported for the similar @xmath0-(bedt - ttf)@xmath1x salts @xcite , which are also known for a strong mott instability @xcite . while the temperature dependence of the oscillation amplitude is solely determined by the temperature damping factor @xmath42 , the field dependence is contributed by all three damping factors on the right - hand side of eq.([lk ] ) . the dingle factor taking into account the landau level broadening @xmath33 has the form @xcite : @xmath59 where @xmath60 is the cyclotron frequency and @xmath61 is the dingle temperature . the dingle temperature is often associated with the scattering rate @xmath62 @xcite : @xmath63 . finally , the mb factors for the @xmath24 and @xmath29 oscillations are determined , respectively , by the probability amplitudes of bragg reflection or tunneling at the mb junctions @xcite and can be expressed as @xmath64 \label{rmba}\ ] ] and @xmath65 where @xmath66 is the characteristic mb field . equations ( [ lk ] ) - ( [ rmbb ] ) can be used for analyzing the field dependence of the oscillation amplitudes . fig.[ab ] shows the amplitudes of both oscillatory components plotted against inverse magnetic field . the data points are obtained from fft spectra taken in 3t - wide field windows . the height of the @xmath28 peak becomes comparable with the fft background noise level above 20 t ; therefore , for this frequency only the data in the field range 12.8 to 20 t is presented . oscillations as a function of inverse magnetic field ; ( b ) the same for the @xmath29 oscillations . both amplitudes are measured in the same units . the lines are fits by eqs.([lk])-([rmbb ] ) , see text.,scaledwidth=45.0% ] one readily sees from eqs.([rd ] ) and ( [ rmbb ] ) that @xmath67 has the same functional dependence on @xmath68 as the dingle factor . therefore , the dingle temperature and mb field can not be separately determined from the field dependence of the @xmath29 oscillations , which dominate our sdh spectrum . on the other hand , the @xmath68-dependence of @xmath69 is rather weak as compared to the exponential behaviour of @xmath70 and @xmath71 . this limits the accuracy of our analysis . nevertheless , we will show that it enables us to draw some important qualitative conclusions . for fitting we first assume that scattering is momentum - independent and , thus , the dingle temperature is the same for both orbits . in this case , substituting the cyclotron masses determined above and performing an iterative fit of both @xmath72 and @xmath73 , we obtain @xmath74k and @xmath75 t . taking into account that , despite the heavier mass , the @xmath29 oscillations strongly dominate the sdh spectrum even at @xmath76 t , the obtained mb field seems to be much too high . indeed , in order to match the observed relation between the amplitudes , we have to assume an unreasonably large ratio of the prefactors in eq.([lk ] ) : @xmath77 . since the @xmath24 orbit constitutes approximately one half of the @xmath29 orbit ( see inset in fig.[fft ] ) , one would expect this ratio to be @xmath78 . of course , the ratio @xmath79 , being determined by the inplane momentum dependence of the fermi surface parameters , needs not be exactly 2 . nevertheless , a modification by more than two orders of magnitude looks highly unlikely . a key to resolving the apparent controversy is in a proper consideration of the influence of many body interactions . if the latter depend on the inplane momentum , the above assumption of a common dingle temperature for both orbits is no longer valid . although the rigorous analysis is very difficult , fortunately , the consideration may be strongly simplified by taking into account compensation of the renormalization effects on the cyclotron mass and dingle temperature . for electron - phonon interactions such compensation was observed long ago on hg @xcite : the dingle factor @xmath43 was found to be independent of temperature despite a considerable @xmath80-dependence of the electron - phonon scattering ( see also ref . for a review ) . it has been shown theoretically @xcite that in a large field and temperature range the product @xmath81 can be approximated by the product of the bare cyclotron mass @xmath82 and dingle temperature @xmath83 in the absence of electron - phonon scattering . while the case of electron - electron interactions is less studied in this respect , martin et al . @xcite have proposed that the same compensation should hold for any inelastic processes , including the electron - electron scattering . turning to our compound , we note that the @xmath24 pocket of the fermi surface contains extended flat segments ( see inset in fig.[fft ] ) and hence has a pronounced nesting property . one can expect that electron correlations are particularly enhanced at the nesting wave vector . this is probably the reason for the unusually high effective cyclotron mass ratio @xmath84 as compared to the value @xmath85 typical of other @xmath0-type organic conductors @xcite . indeed , the other salts do not have such flat segments and the interactions are believed to be momentum independent @xcite . we , therefore , redo the analysis of the oscillations amplitudes in fig.[ab ] , substituting the bare cyclotron masses @xmath86 and @xmath87 in the dingle factor @xcite . with an additional condition that the prefactor ratio @xmath88 , the fit yields a reasonably small value of the mb field , @xmath89mt and only slightly different dingle temperature values , 3.1k and 3.5k for the @xmath24 and @xmath29 oscillations , respectively . of course , the present analysis is far from being precise . an improvement can be achieved when more reliable band mass estimations based on advanced , first - principle band structure calculations are available . further theoretical and experimental studies are required for evaluation of the fermi surface properties entering the prefactors @xmath90 in eq.([lk ] ) for the oscillation amplitude . nevertheless , already now we can conclude that the behaviour of the sdh oscillations clearly reveals the importance of electron correlations in the present organic conductor . we are grateful to s.m . winter and p.d . grigoriev for useful discussions . the work was supported by the german research foundation ( dfg ) via the grant ka 1652/4 - 1 . the high - field measurements were done under support of the lncmi - cnrs , member of the european magnetic field laboratory ( emfl ) . v.n.z . acknowledges the support from rfbr grant 15 - 02 - 02723 .	we present magnetoresistance studies of the quasi - two - dimensional organic conductor @xmath0-(bets)@xmath1mn[n(cn)@xmath1]@xmath3 , where bets stands for bis(ethylenedithio)tetraselenafulvalene . under a moderate pressure of 1.4kbar , required for stabilizing the metallic ground state , shubnikov - de haas oscillations , associated with a classical and a magnetic - breakdown cyclotron orbits on the cylindrical fermi surface , have been found at fields above 10 t . the effective cyclotron masses evaluated from the temperature dependence of the oscillation amplitudes reveal strong renormalization due to many - body interactions . the analysis of the relative strength of the oscillations corresponding to the different orbits and its dependence on magnetic field suggests an enhanced role of electron - electron interactions on flat parts of the fermi surface .
thanks to the growing research effort on the magnetoelectric compounds - which below a critical temperature simultaneously display ferroelectricity and some kind of magnetic order - several mechanisms that can couple magnetism to a macroscopic polarization have been identified . among them , one of the most interesting is the dzyaloshinsky - moriya ( dm ) effect @xcite , which gives rise to helical spin structures through an exchange hamiltonian which depends on the vector product between adjacent spins . such magnetism can also be associated with an electric polarization @xmath9 whose expression contains the same vectorial term @xcite . the compound here studied , ba@xmath0cuge@xmath0o@xmath1 ( bcgo ) , is the only member of ba@xmath0xge@xmath0o@xmath1 family which develops helical magnetism at liquid helium temperatures via the dm mechanism @xcite . indeed , for @xmath10=mn ( spin @xmath11 = 5/2 ) and co ( @xmath11 = 3/2 ) the insulator is antiferromagnetic ( af ) below nel temperatures @xmath2 = 4.0 and 6.7 k , respectively @xcite . below @xmath2 these systems develop a magnetoelectricity due to the spin - dependent hybridization between the @xmath12 orbitals of the transition metal and the @xmath13 orbitals of oxygen @xcite . ba@xmath0cuge@xmath0o@xmath1 instead , having the cu@xmath14 ion @xmath11 = 1/2 , is magnetically much less anisotropic . below @xmath2 = 3.2 k , it thus displays a quasi - af cycloidal , incommensurate magnetism . however , despite the absence of a center of inversion symmetry in the crystal structure , bcgo does not develop spontaneous ferroelectricity @xcite . nevertheless , a macroscopic electric polarization can be induced in it by an external magnetic field @xcite . as the magnetoelectric properties of ba@xmath0cuge@xmath0o@xmath1 have been extensively discussed in refs . , the present work is focused on its lattice dynamics , both from an experimental and from a theoretical point of view . we have thus measured the reflectivity @xmath15 of a single crystal of this compound , from 80 to 6000 @xmath16 and from 7 to 300 k , with the radiation polarized both along the @xmath17 ( or @xmath18 ) axis and along the @xmath4 axis of the tetragonal unit cell . for both polarizations , the real part @xmath19 of the optical conductivity extracted from @xmath15 displays in the far infrared a number of well defined phonon peaks , as ba@xmath0cuge@xmath0o@xmath1 is an excellent insulator . the measured frequencies and intensities have been compared with the predictions of shell - model calculations , which started from the structural determinations by x - ray diffraction . bcgo was found @xcite to crystallize in the non - centrosymmetric tetragonal space group p42@xmath20 m , with lattice parameters @xmath21 = 0.8466 nm and @xmath4 = 0.5445 nm these values are presumably measured at room temperature , and no structural changes induced by temperature are reported in the literature at best of our knowledge . a schematic view of its unit cell - corresponding to two formula units - is reported in fig . [ structure ] , which reproduces that reported in ref . . the ba@xmath14 planes , orthogonal to the @xmath4 axis , separate the layers made of corner - sharing geo@xmath22 and cuo@xmath22 tetrahedra . it is in the resulting two - dimensional square lattice of cu@xmath14 ions , that the cu spins interact through the dm mechanism and the helical magnetic structure takes place . single crystals of ba@xmath0cuge@xmath0o@xmath1 were grown by the floating zone technique at a pressure of 3 bar in oxygen gas atmosphere with the growth speed of 0.5 mm / hour @xcite . the morphology , phase composition and purity of the grown crystals were inspected by high resolution x - ray diffraction and scanning electron microscopy combined with energy dispersive spectroscopy . the crystal orientation was determined by x - ray back - reflection laue method showing well defined , neither distorted nor smeared - out spots . moreover , the crystals used in this work were cut in such a way that the largest surface contained the @xmath17 ( or @xmath23 ) and the @xmath4- axes . this surface was finely polished with polycrystalline diamond suspension down to 0.3 @xmath24 m in grain size while the orientation of the @xmath25 ( or @xmath26 ) axes with respect to the sample edges was assessed by electron back - scattered kikuchi diffraction technique . the dimensions of the crystals employed in this work were approximately 3@xmath272@xmath271 mm . in the visible range they look transparent and colourless like a piece of glass . the reflectivity @xmath15 of this surface was measured with a rapid - scanning michelson interferometer at a resolution of 2 @xmath16 . two different polarizations of the radiation field were used , one along the @xmath4 axis and the other one perpendicular to the @xmath4 axis ( the latter being indicated , in the following , as belonging to the @xmath3 plane ) . the sample was mounted on the cold finger of a helium - flow cryostat . even if the phonon region is limited to 850 @xmath16 , @xmath15 was measured by a rapid - scanning interferometer up to 6000 @xmath16 in order to perform on it accurate kramers - kronig ( kk ) transformations . the reference was a gold film evaporated in - situ onto the sample . as the crystal is nearly transparent in the midinfrared , to avoid interference fringes its back surface was left unpolished and wedged @xcite with respect to the front surface . the real part @xmath19 of the optical conductivity was then extracted from @xmath15 by standard kk transforms . the reflectivity was extrapolated to @xmath28 by an inverse power law of @xmath29 , and to @xmath30 by lorentzian fits . the raw reflectivity data of ba@xmath0cuge@xmath0o@xmath1 are shown in fig . [ rfir ] in the frequency range of interest here , for the radiation polarized both in the @xmath3 plane ( a ) and along the @xmath4 axis ( b ) , either at 300 k and 7 k. no interesting effects are shown by the spectra at intermediate temperatures . both spectra look very similar to the corresponding reflectivity spectra of ba@xmath0coge@xmath0o@xmath1 , as reported in the same frequency range in ref . . they are typical of an insulating crystal , with a comb of phonon lines in the far infrared , and a flat and very low reflectivity ( 6 - 7 % ) in the midinfrared range ( not shown ) . the optical conductivity extracted from the data of fig . [ rfir ] is shown in fig . [ sigma_ab ] in the far - infrared region , both at the lowest and highest measured temperatures . we have detected fourteen transverse optical ( to ) phonon modes with the radiation polarized in the @xmath3 plane , out of the eighteen @xmath5 modes predicted by a factor - group analysis of the p42@xmath20 m unit cell , and all the ten @xmath6 modes predicted along the @xmath4 axis . the absence of line frequencies common to both polarizations confirms that the polarizer was correctly oriented and that the sample was a single crystal . in order to understand the complex phonon spectrum of bcgo , we have used a shell model ( sm ) which enabled us to perform lattice dynamical calculations and to compare the results with the measured @xmath31 . the sm has been successfully applied to different compounds ( oxides and hydrides ) where the effects of the anion polarizability turned out to be important for a proper description of the vibrational properties.@xcite the ionic polarizability is taken into account by considering electronic shells with charge @xmath32 coupled harmonically by a force constant @xmath33 to an atomic core . the sm includes long - range coulomb interactions between all charged shells and cores and shell - shell short - range interactions arising from the wavefunction overlap between neighboring ions . we have considered short - range interactions of the born - mayer type , @xmath34 , for the cu - o , ba - o and ge - o bonds . the lattice constants and the atomic positions were taken from the experimental data of ref . . to calculate the oscillator frequencies and strengths we have fixed the total charges @xmath35 of the ions to their nominal values , _ , @xmath36 = -2 , @xmath37 = 2 , @xmath38 = 2 , and @xmath39 = 4 . the model contains 14 adjustable parameters . the initial parameter values were taken from refs . and were subsequently modified in order to fit the measured infrared phonon frequencies at the brillouin zone center . the calculations were carried out with the help of the gulp code.@xcite due to the complexity of the structure and consequently of the developed sm , the use of automatic searches for parameters included in the code turned out to be ineffective.@xcite thus , the adjustment of the parameters was made by hand . the final set of sm parameters which gave the best fit to data is shown in table i. the theoretical frequencies @xmath40[th ] of the transverse optical ( to ) infrared phonons obtained with the model are shown in table ii for the @xmath3 plane and in table iii for the @xmath4 axis . also shown are the calculated oscillator strengths @xmath41[th ] for each @xmath42-th to infrared mode.@xcite the resulting components of the dielectric function tensor , which here is diagonal and also symmetric in the @xmath3 plane , are reported in table iv in both limits of zero frequency and high frequency . in tables ii and iii we have also reported the corresponding values obtained by fitting to the experimental optical conductivity the lorentzian model @xmath43 in eq . [ sigma_fit ] , @xmath44 is measured in @xmath45 @xmath16 , while @xmath46 and @xmath47 are the central frequency , and the linewidth of the @xmath42-th transverse optical mode , respectively , in @xmath16 . @xmath48 is the oscillator strength in @xmath49 . + .2 cm [ cols="^,^,^,^,^,^",options="header " , ] cccccccccc phonon ( @xmath50 ) ] & @xmath40[th ] & @xmath40[7 k ] & @xmath40[300 k ] & @xmath41[th ] & @xmath41[7 k ] & @xmath41[300 k ] & @xmath51[7 k ] & @xmath51[300 k ] + 1 & 59 & & & 13400 & & & & + 2 & 75 & 84 & 82 & 6700 & 27000 & 12500 & 2 & 5 + 3 & 108 & 103 & 102 & 1700 & 8000 & 2300 & 2 & 3 + 4 & 129 & & & 11800 & & & & + 5 & 142 & 152 & 146 & 3500 & 14500 & 9500 & 7.5 & 15 + 6 & 179 & 187 & 182 & 2600 & 30000 & 26500 & 4 & 18 + 7 & 224 & 217 & 212 & 8750 & 26000 & 19000 & 4 & 11 + 8 & 258 & 257 & 258 & 970 & 33000 & 32000 & 21 & 25 + 9 & 261 & 274 & 270 & 11100 & 40000 & 22600 & 6 & 15 + 10 & 319 & 310 & 305 & 126000 & 89000 & 39000 & 11 & 20 + 11 & 334 & 315 & 313 & 46000 & 111000 & 120000 & 4 & 15 + 12 & 374 & 367 & 371 & 1700 & 26000 & 17000 & 8 & 30 + 13 & 443 & & & 30 & & & & + 14 & 489 & & & 450 & & & + 15 & 766 & 710 & 707 & 21500 & 213000 & 153000 & 7 & 17 + 16 & 776 & 714 & 724 & 64000 & 64000 & 42000 & 10 & 23 + 17 & 786 & 772 & 772 & 2500 & 5300 & 3000 & 8 & 8 + 18 & 836 & 844 & 838 & 16500 & 24000 & 12000 & 7 & 10 + [ table ii ] ccccccccc phonon ( @xmath50 ) & @xmath40[th ] & @xmath40[7 k ] & @xmath40[300 k ] & @xmath41[th ] & @xmath41[7 k ] & @xmath41[300 k ] & @xmath51[7 k ] & @xmath51[300 k ] + 1 & 137 & 109 & 100 & 19400 & 17000 & 24000 & 7 & 22 + 2 & 157 & 130 & 131 & 5600 & 6500 & 7500 & 5 & 10 + 3 & 198 & 147 & 146 & 43000 & 121000 & 66000 & 3 & 16 + 4 & 267 & 278 & 269 & 1600 & 1900 & 3300 & 4 & 19 + 5 & 318 & 321 & 320 & 100 & 700 & 300 & 10 & 12 + 6 & 412 & 390 & 391 & 21600 & 58000 & 30000 & 15 & 19 + & & 448 & 446 & & 36000 & 43000 & 11 & 27 + 7 & 488 & 488 & 485 & 56400 & 74000 & 47000 & 9 & 18 + 8 & 559 & & & 8500 & & & & + 9 & 781 & 775 & 771 & 32000 & 174000 & 91000 & 7 & 9 + 10 & 794 & 791 & 786 & 41800 & 84000 & 81000 & 17 & 21 + [ table iii ] cccc @xmath52 & @xmath53 & @xmath54 & @xmath55 + 10.34 & 4.67 & 1.96 & 1.77 + [ table iv ] as one can see in tables ii and iii as well as in figs . [ sigma_ab ] and [ sigma_c ] , there is a good agreement - within the experimental linewidths - between the observations and the theoretical predictions , as far as the phonon frequencies are concerned . the main discrepancies concern phonons 15 and 16 of the @xmath3 plane and modes 1 - 3 of the @xmath4 axis , where the calculated frequencies are much higher than the observed ones . this may be ascribed either to anharmonic effects which particularly affect those modes , not taken into account in the model , or to an overestimation of the shell - shell repulsive interactions . the latter problem may be specifically present in the cu - o and/or ge - o bonds in the case of the ab - polarized phonons 15 and 16 , and in the ba - o bond in the case of the c - polarized modes 1 - 3 . concerning the oscillator strengths , most calculated values reported in the same tables and figures match the order of magnitude of those measured in the real system , but are systematically lower . this may be mainly due to the so called `` charged phonon '' effect@xcite by which phonons acquire an anomalously large spectral weight due to interatomic movement of charge associated with their displacement pattern . such effect is not included in the shell model which only takes into account on - site distortions of the atomic cloud . this results in a transfer of spectral weight to the far infrared region from the electronic bands in the near infrared and the visible . similar `` dressed phonons '' were observed in several perovskites with high polarizability @xcite through a strong increase of the phonon intensities at low temperature , especially for the lowest - frequency modes , and through the failure of the @xmath7-sum rule @xmath56 when , as in eq . [ sum ] , it is restricted to the far infrared ( fir ) . indeed , in ba@xmath0cuge@xmath0o@xmath1 both the o and cu ions are largely polarizable as shown by the large shell charges and the small core - shell force constants ( see table i ) required by the model fitting of the observed spectra . we therefore compared with each other the intensities provided by the fits at the lowest and highest temperatures of the experiment . as shown in tables ii and iii , most phonons are more intense at low temperature , so that @xmath57[7 k ] is larger than @xmath57[300 k ] by about 40 @xmath58 for the @xmath3 plane and 45 @xmath58 for the @xmath4 axis . this effect , which `` dresses '' the phonons , is not taken into account in the shell model here employed , and may explain most of the discrepancies between theory and experiment which emerge in tables ii and iii . the @xmath7-sum - rule violation has also the interesting implication that the dielectric constant of ba@xmath0cuge@xmath0o@xmath1 , which is related to @xmath57 @xcite , is higher than the theoretical values in table iv and should also considerably increase when cooling the system . as the lowest temperature here reached is larger than @xmath2 , this effect would not be related to the helical spin ordering , but to a redistribution of the charges in the unit cell . the atomic displacements for selected phonons of the @xmath3 plane are shown in fig.[displ_ab ] . they are labeled by their number in table ii and by the theoretical frequency in @xmath16 . surprisingly , we find that in many modes the cuo tetrahedra and the geo tetrahedra have a similar pattern , and this is the case of the four modes shown in fig.[displ_ab ] . this behavior is probably due to the similarity of the ge and the cu mass which makes the tetrahedral `` molecules '' to vibrate at similar frequencies and to mix appreciably their modes in the solid . therefore , we shall discuss the pattern by referring to the tetrahedra without specifying the central atom . in the 129 @xmath16 mode ( a ) the atom at the center of the tetrahedra and three vertex oxygens move approximately in the same direction while a fourth vertex oxygen moves in a quasi - perpendicular direction . thus the angle of only one cu - o ( or ge - o ) bond is significantly perturbed . in the 224 @xmath16 mode ( b ) two oxygens move approximately parallel to the central atom and two others move in a quasi - opposite direction , causing a larger distortion of the tetrahedra even if , again , mainly in the internal angles . in the 319 @xmath16 mode ( c ) the pattern is again similar for the cu and ge tetrahedra but they move out of phase . the central cu moves toward a vertex o when the central ge moves away from a vertex o. as the other atoms move perpendicular to the bonds , they mainly change their angles , causing the mode to be a mixture of bending and stretching . this explains its relatively large energy . the 776 @xmath16 mode ( d ) has a stronger stretching character as shown by its high energy , although one never finds in bcgo either purely stretching or purely bending modes , as it is the case in higher - symmetry solids . this also explains why here , at variance with many oxides where the highest vibrational frequency corresponds to a pure oxygen stretching mode , the highest - energy lines do not shift appreciably when decreasing the temperature . finally , fig . [ displ_c ] shows selected phonon modes polarized along the @xmath4 axis . they are labeled by their number in table iii and theoretical frequency in @xmath16 . mode ( a ) at 137 @xmath16 shows a rigid motion of the cu tetrahedra with the ba in antiphase . the ge ions practically do not move and act as `` nodes '' of the vibration . in most other modes , instead , the two kinds of tetrahedra show very similar displacements , as discussed above for the @xmath3 plane . in mode ( b ) at 198 @xmath16 the cu and ge tetrahedral `` molecules '' move in antiphase and the ba ions have smaller displacements . mode ( c ) at 488 @xmath16 is a mixture of stretching and bending , similar to mode ( c ) of fig . 5 . finally , in mode ( d ) at 794 @xmath16 , the stretching character of the bonds bridging the o s to the center of the tetrahedra prevails . in the present work we have studied the lattice dynamics of ba@xmath0cuge@xmath0o@xmath1 , by infrared reflectivity measurements with polarized radiation , down to temperatures close to those where a helimagnetic phase takes place via the dzyaloshinsky - moriya mechanism . the number of the observed phonon lines is lower ( for the @xmath3 plane ) or equal ( for the @xmath4 axis ) to that predicted for the p42@xmath20 m cell of ba@xmath0cuge@xmath0o@xmath1 , and no linesplitting has been observed when cooling the sample to 7 k. therefore , our spectra confirm that the tetragonal symmetry is conserved down to the lowest temperatures , with no appreciable orthorhombic distortion . the optical conductivity extracted from @xmath15 has been fit by a sum of lorentzians , and their parameters have been compared with the results of shell - model calculations . these have correctly predicted the observed frequencies , within the experimental linewidths , except for a few modes where the theoretical values are systematically higher . the discrepancy may be due either to anharmonic effects , not taken into account in the model , or to an overestimation of the shell - shell repulsive interactions . a systematic underestimation with respect to the observed values is instead exhibited by the calculated oscillator strengths . we have tentatively explained this effect by considering that `` charged - phonon '' effects - not considered in the model - can increase the dipole moment of those vibrations , due to the distortion of the electron clouds going beyond one atom . such interpretation is consistent both with the strong increase observed in the phonon intensities at low temperature , and with the failure of the optical sum rule when it is restricted to the phonon region . although we can not make a quantitative estimate , because of lack of experimental information in the terahertz region and below , the remarkable increase in the phonon intensities that we observe at temperatures which are low , but higher than @xmath2 , suggests that the dielectric constant of ba@xmath0cuge@xmath0o@xmath1 may increase at those temperatures , for a redistribution of the electric charge within the cell . measurements of the static dielectric constant could verify this expectation . we thank paolo barone for suggesting software tools useful to the figure preparation . sk thanks j. gale and r. migoni for useful discussions , and acknowledges support from the consejo nacional de investigaciones cientficas y tcnicas de la repblica argentina . 99 i. e. dzyaloshinskii , j. phys . chem . solids * 4 * , 241(1958 ) . t. moriya , phys . rev . * 120 * , 91 ( 1960 ) . h. murakawa , y . onose , s. miyahara , n. furukawa , and y. tokura , phys . b * 85 * , 174106 ( 2012 ) . a. zheludev , g. shirane , y. sasago , n. kiode , and k. uchinokura phys . b * 54 * , 15163 ( 1996 ) ; a. zheludev , s. maslov , g. shirane , y. sasago , n. kiode , and k. uchinokura phys . b * 57 * , 2968 ( 1998 ) ; a. zheludev , s. maslov , g. shirane , i. tsukada , t. masuda , k. uchinokura , i. zaliznyak , r. erwin , and l. p. regnault , phys . rev . b * 59 * , 11432 ( 1999 ) . j. chovan , m. marder , and n. papanicolaou , phys . b * 88 * , 064421 ( 2013 ) . t. masuda , s. kitaoka , s. takamizawa , n. metoki , k. kaneko , k. c. rule , k. kiefer , h. manaka , and h. nojiri , phys . b * 81 * , 100402(r ) ( 2010 ) . a. zheludev , t. sato , t. masuda , k. uchinokura , g. shirane , and b. roessli , phys . b * 68 * , 024428 ( 2003 ) . h. murakawa , y . onose , s. miyahara , n. furukawa , and y. tokura , phys . lett . * 105 * , 137202 ( 2010 ) . k. yamauchi , p. barone , and s. picozzi , phys . b * 84 * , 165137 ( 2011 ) . h. murakawa , y. onose , and y. tokura , phys . 103 , 147201 ( 2009 ) . m. tovar , r.e . dinnebier , w. eysel , mater . forum * * 2*78 - 281 , 750 ( 1998 ) . r. fittipaldi , g. balakrishnan , and a. vecchione , preprint submitted to j. cryst . growth . k. momma and f. izumi , j. appl . crystallogr . * 4*4 , 1272 - 1276 ( 2011 ) . p. dore , a. nucara , d. cannav , g. de marzi , p. calvani , a. marcelli , r. s. sussmann , a. j. whitehead , c. n. dodge , a. j. krehan , and h. j. peters , applied optics * 3*7 , 5731 , 1998 . v. hutanu , a. sazonov , m. meven , h. murakawa , y. tokura , s. bordcs , i. kzsmrki , and b. nfrdi , 86 * , 104401 ( 2012 ) . s. koval , r. migoni and h. bonadeo , j. phys . : condens . matter * 4 * , 4759 ( 1992 ) . s. koval , m. g. stachiotti , r. l. migoni , m. s. moreno , r. c. mercader , and e. l. peltzer y blanc , phys . rev . b * 54 * , 7151 ( 1996 ) . j. lasave , f. dominguez , s. koval , m. g. stachiotti and r. l. migoni , j. phys . : condens . matter * 17 * , 7133 ( 2005 ) . j. lasave , j. kohanoff , r.l . migoni , and s. koval , physica b * 404 * , 2736 ( 2009 ) . m. braden , w. reichardt , b. hennion , g. dhalenne , and a. revcolevschi , phys . b * 66 * , 214417 ( 2002 ) . s. tinte , m. g. stachiotti , m. sepliarsky , r. l. migoni , and c. o. rodriguez , j. phys . : condens . matter * 11 * , 9679 ( 1999 ) . j.d . gale , j. chem . faraday trans . * 1 * , 629 ( 1997 ) . s. koval and r. migoni , phys . rev . b * 51 * , 6634 ( 1995 ) . m. j. rice , solid state commun . 31 , 93 ( 1979 ) . c. c. homes , t. vogt , s. m. shapiro , s. wakimoto , and a. p. ramirez , science * 293 * , 673 ( 2001 ) . c. c. homes , t. vogt , s. m. shapiro , s. wakimoto , m. a. subramanian , and a. p. ramirez , phys . b * 67 * , 092106 ( 2003 ) .	the lattice dynamics of ba@xmath0cuge@xmath0o@xmath1 , a compound which develops dzyaloshinsky - moriya ( dm ) helical magnetism below @xmath2 = 3.2 k , has been studied by measuring the infrared reflectivity of a single crystal with the radiation polarized both in the @xmath3 plane and along the @xmath4 axis of its tetragonal cell , from 7 k to 300 k. in this compound , where the unit cell has no inversion symmetry , fourteen @xmath5 phonon modes of the @xmath3 plane , out of the eighteen predicted , and all the ten @xmath6 modes of the @xmath4 axis , have been observed . they have been assigned to the atomic motions by a comparison with shell - model calculations , which provided vibrational frequencies in good agreement with the experiment , while most calculated intensities turned to be much lower than the experimental values . this discrepancy has been tentatively explained by assuming strong electron - phonon interactions , a hypothesis supported by the failure of the @xmath7- sum rule if restricted to the phonon region . indeed , we observe a remarkable increase in the oscillator strengths at @xmath8 s low but higher than @xmath2 , which suggests that the dielectric constant of ba@xmath0cuge@xmath0o@xmath1 may increase at those temperatures .
the braneworld models became an active research area in high energy physics in the last years . the seminal works of randall - sundrum ( rs ) @xcite introduced an infinite extra dimension by means of a warped compactification . thereafter , several models explored both the field @xcite as cosmological properties @xcite of the rs models . in six dimensions , by assuming a static and axial symmetry for the bulk , the geometry of the brane is similar to cosmic string space - time . then , these models are called string - like braneworlds @xcite . even though the string - like models have the advantage of traping free gauge fields @xcite and minimally coupled dirac fermions @xcite , they present some issues about the core of the source . the gherghetta - shaposhnikov ( gs ) model @xcite , for instance , does not satisfy the metric regularity conditions at the brane nor the dominant energy condition @xcite . this is due to the metric proposed be only a vacuum solution of the einstein equations . the gs model can be regarded as an infinitely thin string - like model . in order to study the source properties and to suppress the anomalies of the string - like models , giovannini _ et al _ proposed a braneworld scenario constructed from an abelian vortex @xcite . the solution found numerically satisfies all the regularity and energy conditions , but the analytical solution is still lacking @xcite . moreover , the same solutions leading to gravity localization , also lead to the localization of the gauge zero mode @xcite . afterward , de carlos - moreno proposed a supersymmetric model without bulk cosmological constant free of regularity problems @xcite , whereas papantonopoulos _ et al _ regularized the conical behavior near the brane by adding a ring - like structure at the brane @xcite . more recently , silva - almeida proposed a resolution scheme based in an effective conifold transition in the internal space @xcite . the resolution of the conical singularity at the core of the brane provides a geometrical flow which smoothes the kaluza - klein ( kk ) modes for the scalar , gauge and dirac fields @xcite . an analytical smooth string - like model was proposed and its gravitational kk modes were studied @xcite . this thick solution extends the gs model being an interior and exterior string - like model . the near brane corrections to the geometry makes the model satisfy all the regularity and energy conditions @xcite . since this string - like model is built from a warped product between a @xmath0brane and a particular steady solution of the ricci flow , called hamilton cigar soliton , the scenario is called the string - cigar model @xcite . the evolution parameter of the ricci flow yields to a varying bulk cosmological constant and changes the ratio among the components of the stress - energy tensor @xcite . moreover , the near brane correction provides a potential well around the brane for the gravitational kk modes @xcite . once defined and studied this smoothed string - like braneworld , it is worth to analyse the behavior of the standard model fields on this scenario . in this article , we study the features of the gauge vector field minimally coupled to this geometry . we show that the radial component has a richer dependence compared with the thin string - like solution presented on ref . @xcite . moreover , the kk tower has a trapped massless @xmath1wave state shifted from the origin due to the core behavior . the displacement of the massless mode happens because the brane - core is shifted from the origin , as in the abelian vortex solution for higher winding numbers @xcite . nevertheless , the source and the massless mode approach to the origin as the geometric evolution parameter increases . the limit of large values of this parameter matches with the thin string results . the dynamics for the massive modes , presented in a sturm - liouville equation @xcite , is quite complex to be studied analytically . then , we accomplish a numerical analysis to find the kk spectrum and the respective eingenfunctions . it turns out that exists a mass gap between the massless and the massive modes for both the thin string - like and string - cigar models . moreover , the known linearly increasing kk masses are obtained . the kk eingenfunctions are all smooth near the brane and they recover the usual string - like pattern asymptotically @xcite . the string - cigar core enhances the amplitude of the kk modes near the brane . performing the analogue schrdinger approach , it turns out that the geometric flow , provided by a geometric parameter , controls the high of the barrier and the width of the potential well . this article is organized as follows : in the sec . [ sec - thestringcigarscenario ] , we review the properties of the thin string - like and string - cigar models as well as the changes provided by the ricci flow on the brane core . in the sec . [ gaugelocalization ] , we investigate the gauge vector massless field and the kk spectrum and study the influence of the geometric flow in this scenario for the @xmath1wave state . further in this section , we analyse the analogue schrdinger potential behavior upon this flow . moreover , some conclusions and perspectives are outlined in sec . [ sec - conclusion ] . in this section , we briefly review the construction of the string - cigar model @xcite . a complete description of the model , as well further details are presented in ref . firstly , let us define a string - like spacetime . let @xmath2 a spacetime that can be split as @xmath3 , where @xmath4 is a @xmath0brane embedded in @xmath2 and @xmath5 is a two dimensional riemannian space . a string - like static braneworld is a @xmath2 with axial symmetry . a warped metric ansatz for this axisymmetric spacetime is @xcite @xmath6 where @xmath7 is the induced metric brane and @xmath8 and @xmath9 are the so - called warp factors . in order to the brane possess a regular geometry , we assume that the warp functions satisfy the regularity conditions , namely @xcite @xmath10 where the prime @xmath11 denotes the derivative according to @xmath12 variable . an example of a string - like geometry is the gs model @xcite , where @xmath13 is the two - dimensional disk of radius @xmath14 and the metric is given by the components @xcite @xmath15 where @xmath16 is a constant . the constant @xmath17 is related to the bulk cosmological constant @xmath18 by @xcite @xmath19 therefore , the gs model is a @xmath20 space time which geometry is an exterior solution for a thin string - like brane @xcite . in the gs model , the relation between the bulk planck mass @xmath21 and the brane planck mass @xmath22 is given by @xcite @xmath23 then , in order to solve the hierarchy problem , gs model imposes a fine tuning between the bulk cosmological constant and the angular tension @xmath24 , what yields to @xcite @xmath25 the inequality ( [ cosmologicalconstantconstrain ] ) constrains the possible values for the constant @xmath17 . in the rs model the curvature is set to be small in order to guarantee the model be obtained from the horava - witten model @xcite . for the gs model , @xmath17 can assume any value satisfying the conditions ( [ clambdarelation ] ) and ( [ cosmologicalconstantconstrain ] ) . despite all these important features , the gs model does not satisfy all the regularity conditions at the origin @xcite . for @xmath26 , instead of a 3-brane we obtain a 4-brane @xcite . further , although the thin brane be flat , the curvature at the origin is non zero @xcite . besides these geometrical issues , the source also does not satisfy all the energy conditions @xcite . the drawbacks of the gs model arise because it is only a vacuum exterior solution of the einstein equations @xcite . in order to solve these problems concerning the thin string - like branes , an interior and exterior solution for a string - like braneworld was proposed taking the cigar soliton as the transverse space @xmath5 @xcite . the cigar soliton is a solution of the geometrical ricci flow whose equation is given by @xcite @xmath27 the ricci flow ( [ ricciflow ] ) defines a family of geometries evolving under a parameter @xmath28 . the metric for the cigar soliton can be written as @xcite @xmath29 the evolution parameter @xmath28 may be identified with the warp constant @xmath17 @xcite , so that , choosing the warp metric components @xmath30 and @xmath31 it defines an axisymmetric braneworld called string - cigar model @xcite . the string - cigar model converges to the gs model far from the origin @xcite . near the origin , the string - cigar geometry presents a conical behavior and it smoothes out the warp factor . as a result , all the regularity conditions are satisfied and thereby , the string - cigar geometry is a smooth interior and exterior string - like solution @xcite . the string - cigar model not only smoothes the gs model near the brane but also provides a geometrical flow due to the variation of the bulk cosmological constant . indeed , the components of the stress - energy tensor evolves under the geometrical flow . for small @xmath17 , the source satisfies the dominant energy condition , whereas for great values @xmath17 , only the weak energy condition is satisfied @xcite . moreover , the width and the position of the brane core also change with the flow . the source approaches to the origin and becomes narrower as @xmath17 increases . therefore , high cosmological constant values situate the source around the origin what agrees with the gs model @xcite . furthermore , the string - cigar model also provides a solution for the hierarchy problem . indeed , the relation between the masses are now given by @xcite @xmath32 therefore , as in the rs model , due to the dominant energy condition and the hierarchy solution , we concern ourselves to the small @xmath17 analysis . it is worthwhile to mention that the near brane corrrections shifts the maximum of the energy density from the origin @xcite . similar results were found by numerical analysis performed by giovannini _ et al _ for higher winding number abelian vortex @xcite . besides , kehagias argued that this displacement due to the conical behavior could be used to explain the cosmological problem @xcite . once we presented how the ricci flow changes the properties of the brane core , we proceed to analyse the effects of this geometrical flow upon the abelian vector field . we begin with a @xmath33 invariant vector field action , namely @xmath34 where @xmath35 . from the action ( [ gaugefieldaction ] ) , the equation of motion is obtained in a straightforward way as @xmath36 let us consider the brane lorentz gauge @xmath37 and a purely radial field configuration , e.g. , @xmath38 , as usual @xcite . in addition , since the @xmath0brane is flat and has an axial symmetry , the radial vector component @xmath39 should not depend on the brane coordinates , i.e. , @xcite @xmath40 using the string - cigar metric and the gauge choice , the equation of motion ( [ gaugefieldmotionequation ] ) takes the form @xmath41 @xmath42 and @xmath43 performing the kaluza - klein decompositions @xcite @xmath44 and @xmath45 the eq . ( [ motionequation1 ] ) yields to @xmath46 thus , in order to match eq . ( [ motionequation1 ] ) with eq . ( [ symmetrygaugecondition ] ) , we restrict ourselves to @xmath1waves states , e.g. , @xmath47 @xcite . further , the eq . ( [ motionequation2 ] ) leads to a general solution to @xmath48 as @xmath49 where @xmath50 is an simple integration constant @xcite . it is worth to mention that the function @xmath48 in the eq . ( [ motionequation7 ] ) is a direct extension to that displayed by oda @xcite . finally , using eq . ( [ motionequation7 ] ) , the eq . ( [ equationmotion3 ] ) turns to be @xmath51 where @xmath52 . ( [ chiequation ] ) governs the behavior of the gauge field through the bulk . it is worth to mention that this equation is rather similar to the graviton radial equation , regardless the change of the factor @xmath53 by @xmath54 @xcite . we now impose the following boundary conditions @xcite @xmath55 which yields the orthogonality relation between @xmath56 and @xmath57 given by @xcite @xmath58 furthermore , it is possible to transform the kk equation ( [ chiequation ] ) into a schrdinger - like equation . consider the change of independent variable @xcite @xmath59 and of dependent variable @xmath60 where @xmath61 , with @xmath62 a constant . the changes of variables ( [ changevariable ] ) and ( [ changeofdependentvariable ] ) transform the eq . ( [ chiequation ] ) into a schrdinger one with @xmath63 fulfilling @xmath64 where @xmath65 is given by @xmath66.\ ] ] the over - dots refers to derivatives with respect to the @xmath67 coordinate . for @xmath68 , a solution for the eq . ( [ chiequation ] ) is the following linear combination : @xmath69 where @xmath70 and @xmath71 are integration constants . since the second function does not satisfy the orthogonality condition ( [ orthogonalityrelation ] ) , we choose the constant function @xmath72 as the solution of eq . ( [ chiequation ] ) . for the thin string - like scenario @xcite , the analogue schrdinger equation reads @xmath73 where @xmath74 . note that the origin in the @xmath12 coordinate is mapped into the origin in the @xmath67 coordinate . the solution of eq.([thinschrodinger ] ) for @xmath75 is given by @xmath76 by making @xmath77 in eq . ( [ thinzeromodeschrodinger ] ) , we obtain a normalizable massless @xmath78 this kk massless mode is trapped at the brane because of the exponential factor . moreover , the amplitude is bigger for the gauge massless mode than for the gravitons @xcite . from the general schrdinger equation ( [ schrodigner ] ) for @xmath68 , the solution satisfying the orthogonality condition ( [ orthogonalityrelation ] ) is given by @xmath79 where @xmath80 is a normalization constant . we plot on figure ( [ fig - masslessmode ] ) the massless mode for the string - cigar and for the thin - string scenarios . these localized modes are responsible for the effective @xmath0brane gauge field . the gauge massless modes are more concentrated at origin when compared with the graviton massless modes @xcite . a worthwhile feature is the displacement of the massless mode from the origin in the string - cigar background . this behavior is also present in the energy density of the model ( see ref . @xcite ) . out of the core , the string - like exponential behavior dominates , whereas at the origin , the near core correction vanishes the mode . matching these two regimes , there is a smooth peak which maximum is around the boundary of the brane core . note that , for great values of the geometric parameter @xmath17 ( and then , the bulk cosmological constant ) , the gauge massless mode in the string - cigar scenario tends to the thin string - like case . using the expressions for the metric factors ( [ warpfunction ] ) and ( [ angularmetric ] ) , we obtain the kk equation from eq . ( [ chiequation ] ) in the form @xmath81\chi_n ' + \operatorname{e}^{(cr-\tanh{cr})}m_n^2\chi_n=0.\ ] ] asymptotically , the eq . ( [ completeequationchi ] ) recovers the thin - string model case , presented in ref . @xcite , as @xmath82 which general solution is @xcite @xmath83 , \label{oda - massive - solution}\ ] ] where @xmath84 are normalization constants and @xmath85 are constant coefficients determined by the boundary conditions . looking at the gravitational case presented in ref . @xcite , the graviton fluctuation @xmath86 has the radial solution @xmath87 , \label{gs - massive - solution}\ ] ] where @xmath88 and @xmath89 are arbitrary constants . thus , the massive modes of the gauge field have a higher amplitude near the brane but they spread less into the bulk when compared with the gravitons . for the thin string - like model , there is a discontinuity between the massive and the massless mode . indeed , by making the limit @xmath90 the massive states converge to @xmath91 not to @xmath92 . further , transforming the eq . ( [ oda - sturmliouville_eq ] ) into a schrdinger - like equation , we find that the kk schrdinger solutions has the form @xmath93,\ ] ] where @xmath94 and @xmath95 are integration constants and @xmath96 . since the kk solutions in eq . ( [ thinmassiveschrodinger ] ) are not defined for @xmath68 , then we can not obtain the massless mode continuously from the massive modes . the existence of this kind of mass gap was also found in negative tension braneworlds in five dimensions @xcite . applying the boundary conditions ( [ boundarycondition ] ) at some cut off point @xmath97 , it is possible to obtain the kk spectrum . in fact , the boundary conditions yield to the system of equations @xmath98 the system above is difficult to be treated analytically . however , for the small mass regime , i.e. @xmath99 , the divergence in the bessel function of second kind yields to @xmath100 . hence , the mass spectrum can be obtained from the equation @xcite @xmath101 then , from the zeroes of the bessel function @xmath102 , where @xmath103 , we obtain the massive spectrum as @xcite @xmath104 note that the gauge masses grow linearly with the discrete bessel function root index @xmath105 , as in the factorizable kaluza - klein model @xcite . therefore , for @xmath106 , the spectrum is discrete and there is an exponential suppressed mass gap between the massless mode and the first massive mode given by @xmath107 once the general kk equation ( @xmath108 ) , as well the boundary condition system ( [ boundaryconditionsystem ] ) , are difficult to solve analytically , we look for approximate solutions by numerical methods . using the matrix method @xcite with second order truncation error , we obtain the complete kk spectrum and eigenfunctions for the gauge field on thin - string and string - cigar scenarios . in order to obtain the gauge field kk spectrum on the thin - string scenario , we approximated the derivatives by finite differences in the grid points @xmath109 with constant step - size @xmath110 . the domain used was @xmath111 $ ] . for the string - cigar scenario , the numerical integration of eq . ( @xmath112 ) was performed on the domain @xmath113 $ ] in order to avoid the singularity at @xmath114 . we plot in the figure [ fig - spectra ] the lowest mass eigenvalues @xmath115 ( @xmath116 ) for both braneworld models . note that the usual linear behaviour from the kaluza - klein theories are reproduced , where the index @xmath50 is the kaluza - klein number . since the eq . @xmath117 is rather complex to be solved analytically , an exact value for @xmath115 can not be obtained to compare with the numerical values . however , once the near - brane correction of the string - cigar braneworld enhances the amplitude of the eigenfunctions @xcite , it is expected the same effect for the spectrum . .,scaledwidth=65.0% ] the eigenfunctions were obtained for both models whose shape is plotted in the figure ( [ fig - autofuncao ] ) for @xmath118 . asymptotically , all the solutions behave as eq . ( [ oda - massive - solution ] ) whereas near the origin the amplitude of the kk modes are greater in the smooth string - cigar scenario than in the thin string - like one . the string - cigar solutions behave as bessel functions of first kind near the origin . scenario.,scaledwidth=50.0% ] scenario.,scaledwidth=50.0% ] the massive states were also investigated numerically from the schrdinger - like equation for the string - cigar scenario . the analogue quantum potential was constructed by numerical interpolation from the numerical integral ( [ changevariable ] ) . we have plotted the potential function @xmath65 on the figure ( [ fig - potential ] ) for some values of the evolution parameter . the potential well has a volcano - shape whose barrier increases and approaches to the origin with the increasing of @xmath17 . the schrdinger - like equation ( [ schrodigner ] ) was solved using the numerov algorithm @xcite . we plotted in the figure ( [ fig - funcaodeonda ] ) two wavefunctions for @xmath119 . the potential well influenced the first cycle of the wavefunctions and , for a moderate mass , the solutions rapidly oscillate . modes with intermediate values of mass smoothly interpolate between the two solutions shown . similar results were found for other values of @xmath17 . for some values of @xmath17 . the thin line is a plot of the potential in the thin string - like background for @xmath120 . , scaledwidth=50.0% ] and for small mass values @xmath121 ( dashed line ) and @xmath122 ( thick line).,scaledwidth=50.0% ] in this article we studied the features of the gauge vector field in the thin string - like and in the thick string - cigar model which evolves under a geometric ricci flow . the analysis was carried out for the @xmath1wave states , i.e. , @xmath47 . the massless mode is localized and smoothed out compared to the massless mode in the thin string - like scenario . the maximum of this mode is displaced from the origin , likewise the stress energy components . asymptotically , the thin - string exponential behavior is recovered , whereas inside the core , the massless mode exhibits a conical behavior . we have obtained the kk spectra for the string - like and string - cigar models by numerical techniques . the well - known linear increasing behaviour was obtained , as well the massive gap between the massless mode and the first massive state in the thin - string background . the massive eigenfunctions present a bigger amplitude near the origin in the string - cigar braneworld whereas asymptotically the kk modes exhibit the usual thin string - like behaviour . thus , the string - cigar geometry provides a near brane correction to the kk modes . we have also obtained numerically the analogue quantum potential . it possess a well known volcano - shape which width of the well and the high of the barrier are controlled by the geometric evolution parameter . as perspectives we intent to obtain the corrections to the coulomb potential @xcite . further , we propose to study the effects of the ricci flow on the dirac fermion field . another interesting subject is the analysis of the massive kk modes for @xmath123 configuration and the search of massive resonant modes . the authors thank the coordenao de aperfeioamento de pessoal de nvel superior ( capes , grant 23038007567/2011 - 46 ) , the conselho nacional de desenvolvimento cientfico e tecnolgico ( cnpq , grant 305766/2012 - 0 ) , and fundao cearense de apoio ao desenvolvimento cientfico e tecnolgico ( funcap , grant bfp-026 - 00368.01.00/13 ) for financial support. we would also like to express our thank to the anonymous referee for their enlightening questions and constructive comments . j. e. g. silva and c. a. s. almeida , phys . d * 84 * , 085027 ( 2011 ) . f. w. v. costa , j. e. g. silva and c. a. s. almeida , phys . d * 87 * , 125010 ( 2013 ) . d. m. dantas , j. e. g. silva and c. a. s. almeida , phys . b * 725 * , 425 ( 2013 ) . j. e. g. silva , v. santos and c. a. s. almeida , class . grav . * 30 * , 025005 ( 2013 ) . r. s. hamilton , three - manifold with positive ricci curvature , j. differential geometry , 17 ( 1982 ) . t.oliynyk , v.suneeta and e.woolgar , nucl . b * 739 * , 441 ( 2006 ) . t. oliynyk , v. suneeta and e. woolgar , phys . d * 76 * , 045001 ( 2007 ) . m. headrick and t. wiseman , class . * 23 * , 6683 ( 2006 ) . n. lashkari and a. maloney , class . * 28 * , 105007 ( 2011 ) . b. chow , p. lu and l. ni , hamilton s ricci flow , science press , 2006 .	in this work , we investigate the properties of the abelian gauge vector field in the background of a string - cigar braneworld . both the thin and thick brane limits are considered . the string - cigar scenario can be regarded as an interior and exterior string - like solution . the source undergoes a geometric ricci flow which is related to a variation of the bulk cosmological constant . the ricci flow changes the width and amplitude of the massless mode at the brane core and recovers the usual string - like behavior at large distances . by means of suitable numerical methods , we attain the kaluza - klein ( kk ) spectrum for the string - like and the string - cigar models . for the string - cigar model , the kk modes are smooth near the brane and their amplitude are enhanced by the brane core . furthermore , the analogue schrdinger potencial is also regulated by the geometric flow .
optic and magneto - optic experiments with graphene layers have been successfully interpreted @xcite so far in a scheme of massless relativistic particles with a conical energy spectrum @xmath0 where @xmath1 is the constant velocity parameter in two bands , @xmath2 , near the k and k points in the brillouin zone . in pure graphene , the chemical potential is situated at the charge neutrality point @xmath3 . however , it can have a nonzero value because of doping or under a gate voltage . thus , the chemical potential is determined by the total number of carriers ( difference of electrons in the upper band and holes in the low band ) @xmath4 where @xmath5 is the fermi function , @xmath6 is the surface of the graphene layer , and the factor 4 takes the valley and spin degeneracy into account . the integration is performed over @xmath7 , the chemical potential is positive for electrons and negative for holes . at the fixed @xmath8 , this condition determines the dependence @xmath9 , shown in fig . [ unscr ] for a relatively low electron concentration . for the conical spectrum , eq . ( [ ed ] ) , the ratio between the kinetic and coulomb energies has a constant value independently of the carrier concentration and the problem of the phase electron - dielectric transition becomes undefined . it was recently discovered @xcite in studying of the shubnikov - de haas oscillations that electron - electron interactions are very important for low carrier concentrations , @xmath10 . while the electron concentration decreases from 10@xmath11 to 10@xmath12 @xmath13 , the velocity parameter @xmath1 grows by three times from its ordinary value 1.05@xmath14 cm / s . the logariphmic renormalization of the velocity for the linear electron dispersion was found by abrikosov and beneslavsky @xcite in the three - dimensional case and in refs . @xcite for two - dimensional graphene . notice , that no phase transition was revealed even at the lowest carrier concentration . we can conclude that coulomb interactions do not create any gap in the spectrum . the renormalized electron dispersion can be written in the form @xmath15\ , , \label{ren}\ ] ] where @xmath16 is the dimensionless electron - electron interaction and @xmath17 describes an effect of a substrate and self - screening in graphene , @xmath18 @xmath19 is the cutoff parameter @xcite . in fig . [ scr ] , we consider a screening effect on the chemical potential at the carrier concentration @xmath20 @xmath13 . equation ( [ ren ] ) is written in the linear approximation in @xmath21 . because the logarithm is assumed to be large , the condition @xmath22 has to be fulfilled , and we suppose this condition in what followed . in this article , we consider the impact of the renormalization on thermodynamic properties of graphene such as the chemical potential , heat capacity , and magnetic moment . for low ( @xmath23 ) and high ( @xmath24 ) temperatures , the analytical expressions for @xmath9 can be obtained from eq . ( [ num ] ) with the renormalization taken into account . for low temperatures , it is convenient to differentiate eq . ( [ num ] ) with respect the temperature , using @xmath25 \left [ -\frac{\partial f\left ( \varepsilon -\mu \right ) } { \partial \epsilon } \right]\,.\end{aligned}\ ] ] here , we have a sharp function of @xmath26 . therefore , in the integrand , the momentum @xmath27/v\ ] ] should be expand near @xmath28 in powers of ( @xmath29 ) , which gives a factor proportional to @xmath30 after the integration . for instance , we get in the case of electron doping @xmath31\left[\mu\frac{d\mu}{dt}+\frac{(\varepsilon-\mu)^2}{t}\right]\\ \times [ 1 - 2g\ln(p_0v/\mu)]d\varepsilon\,,\end{aligned}\ ] ] where we do not differentiate the logarithm because of the condition @xmath32 . integrating , one finds @xmath33 where we denote @xmath34 , positive for electrons and negative for holes . let us notice that this is the known temperature dependence of the chemical potential in the degenerate fermi system at low temperatures . we emphasize that the fermi energy @xmath35 is determined indeed by the carrier concentration @xmath36\ , , \label{mu3}\ ] ] which introduces the renormalization in eq . ( [ mu1 ] ) by means of @xmath35 . for high temperatures , we can expand the integrand in eq . ( [ num ] ) in @xmath37 . introducing the new variable @xmath38 , we get the integral @xmath39 which gives the chemical potential @xmath40\,.\label{high}\ ] ] we see the inverse temperature dependence of the chemical potential , as a collective effect in electron - hole liquids . the renormalization term , correcting the temperature dependence , is presented here explicitly and illustrated in fig . now we consider the electron contribution in the heat capacity . the energy of carriers @xmath41differs from the carrier concentrations , eq . ( [ num ] ) , only by the additional factor @xmath42 in the integrand . therefore , we can follow the same procedure . for low temperatures , @xmath43 , the carrier heat capacity in the case of electron doping writes as @xmath44 \\ \times \left [ -\frac{\partial f\left ( \varepsilon -\mu \right ) } { \partial \varepsilon } \right][1 - 2g\ln(p_0v/\mu ) ] d\varepsilon\\ = \frac{2s}{\pi \left ( \hbar v\right ) ^{2}}\left [ \mu ^{2}\frac{d\mu } { dt}+\frac{2\pi^2}{3}\mu t\right][1 - 2g\ln(p_0v/\mu)]\ , . \end{aligned}\]]using eq . ( [ mu1 ] ) , we have @xmath45\ ] ] in both cases of the electron or hole doping . for high temperatures , @xmath46 , one can perform the expansion of the energy , eq . ( [ eneg ] ) , in the first order of @xmath37@xmath47d\varepsilon\\ & = & \frac{2\pi s \|\mu \|}{3\left ( \hbar v \right ) ^{2}}t^{2}[1 - 2g\ln(p_0v/2t)]\,.\end{aligned}\ ] ] using eq . ( [ high ] ) , we find @xmath48finally,@xmath49 thus , we see that the renormalization modifies the heat capacity at low temperatures , i.e. , in the degenerate statistics . at high temperatures , the heat capacity possesses the constant value and does not reveal any renormalization at least to a first approximation in @xmath50 . magnetic susceptibility is determined by the dependence of the thermodynamic potential on the magnetic field @xmath51 in terms of the electron dispersion @xmath52 for two bands @xmath2 with the landau number @xmath53 we neglect the spin splitting of the levels in comparison with the large landau splitting in graphene . oscillations of the magnetic moment in the semi - classical region can be found applying the poison formula to the thermodynamic potential @xmath54 where the contributions of two bands are written explicitly . the integraton by parts gives @xmath55e^{2\pi ikn}d\varepsilon\ , . \label{sm}\ ] ] for the semi - classical region , we use the quantization rule in the bohr zommerfeld form @xmath56 with the aria enclosed by the electron trajectory for the energy @xmath42 in the momentum space @xmath57 according to eq . ( [ ren ] ) . the main contribution in the integral ( [ sm ] ) comes from the vicinity of the point @xmath58 for the positive and negative @xmath35 , correspondingly . expanding the exponent in the integrand near that points and integrating , one finds @xmath59}{\sinh(2\pi^2 kc\|m(\varepsilon_f)\| t / e\hbar b)}\,,\end{aligned}\ ] ] where @xmath60 is the cyclotron mass . in calculating of the magnetic moment we can derivative only the rapid factor in the argument of @xmath61 with respect @xmath62 : @xmath63}{\sinh[2\pi^2 kc\|m(\varepsilon_f)\| t / e\hbar b]}\,,\end{aligned}\ ] ] where the carrier concentration @xmath64 . this is the standard lifshiz - kosevich formula used in ref . @xcite for the interpretation of experimental data concerning the velocity renormalization . there are two important features : first , the aria @xmath65 and the effective mass @xmath66 should be taken at the renormalized fermi energy corresponding to the carrier concentration and , second , the factor in front of the sum differs from the 3d case since the integration over @xmath67 is absent now . it is interesting to compare the amplitude of oscillations with the monotonic part of the magnetic moment , ref . @xcite , @xmath68 thus , we see that the ration of the oscillating and monotonic parts of the magnetic moment has the order @xmath69}\,.\ ] ] to observe the oscillations , the argument of @xmath70 has to be small or at least on the order of unity . then , the monotonic part of the magnetic moment can be observable only at relatively high temperatures , @xmath71 . it should be emphasize that such a transport property as the electronic conductivity is not sensitive to the electron - electron interaction since the conductivity does not depend indeed on the velocity parameter @xmath1 . the renormalization of the electron spectrum due to coulomb interactions in graphene is noticeable in thermodynamic properties especially at low temperatures and for small carrier concentrations @xmath72 @xmath13 , as can be seen from eqs . ( [ high ] ) , ( [ cap ] ) , and ( [ sus ] ) . however , the interesting temperature dependences @xmath73 for the chemical potential at high temperatures @xmath46 , appears independently of electron - electron interactions . the detection of the renormalization requires the high accuracy in experiments because the renormalization can be concealed by increasing of the velocity parameter @xmath1 [ see eq . ( [ mu3 ] ) and fig . [ scr ] ] . we gratefully acknowledge andrey varlamov for useful discussions . this work was supported by the russian foundation for basic research ( grant no . 13 - 02 - 00244a ) and the simtech program , new centure of superconductivity : ideas , materials and technologies ( grant no .	the impact of renormalization of the electron spectrum on the chemical potential , heat capacity , and oscillating magnetic moment is studied . the cases of low and high temperatures are considered . at low temperatures , doped graphene behaves as the usual fermi liquids with the power temperature laws for thermodynamic properties . however , at high temperatures and relatively low carrier concentrations , it exhibits the collective electron - holes features : the chemical potential tends to its value in the undoped case going with the temperature to the charge neutrality point . simultaneously , the electron contribution into the heat capacity tends to the constant value , as in the case of the boltzmann statistics .
kohn - sham density functional theory ( dft ) @xcite has a relatively high accuracy / cost ratio , which makes it a popular electronic structure method for predicting material properties and behavior . in dft , the system of interacting electrons is replaced with a system of non - interacting electrons moving in an effective potential @xcite . the electronic ground - state in dft is typically determined by solving for the kohn - sham orbitals , the number of which is commensurate with the size of the system , i.e. number of electrons @xcite . since these orbitals need to be orthonormal , the overall solution procedure scales cubically with the number of atoms @xcite . in order to overcome this restrictive scaling , significant research has focused on the development of linear - scaling methods @xcite . nearly all of these approaches , in one form or the other , employ the decay of the density matrix @xcite in conjunction with truncation to achieve linear - scaling @xcite . however , an efficient linear - scaling algorithm for metallic systems at low temperatures still remains an open problem @xcite . orbital - free dft ( of - dft ) represents a simplified version of dft , wherein the electronic kinetic energy is modeled using a functional of the electron density @xcite . commonly used kinetic energy functionals include the thomas - fermi - von weizsacker ( tfw ) @xcite , wang - teter ( wt ) @xcite , and wang , govind & carter ( wgc ) @xcite variants . amongst these , the wt and wgc functionals are designed so as to match the linear - response of a homogeneous electron gas @xcite . previous studies have shown that of - dft is able to provide an accurate description of systems whose electronic structure resembles a free - electron gas , e.g. aluminum and magnesium @xcite . there have been recent efforts to extend the applicability of of - dft to covalently bonded materials @xcite as well as molecular systems @xcite . in essence , of - dft can be viewed as a ` single - orbital ' version of dft , wherein the cubic - scaling bottleneck arising from orthogonalization is no longer applicable . in addition , of - dft possesses an extremely favorable scaling with respect to temperature compared to dft @xcite . overall , of - dft has the potential to enable electronic structure calculations for system sizes that are intractable for dft . the plane - wave basis is attractive for performing of - dft calculations @xcite because of the spectral convergence with increasing basis size and the efficient evaluation of convolutions using the fast fourier transform ( fft ) @xcite . however , developing implementations which can efficiently utilize modern large - scale , distributed - memory computer architectures is particularly challenging . further , evaluation of the electrostatic terms within the plane - wave basis typically scales quadratically with the number of atoms @xcite . in view of this , recent efforts have been directed towards developing real - space approaches for of - dft , including finite - differences @xcite and finite - elements @xcite . amongst these , the finite - element method provides the flexibility of an adaptive discretization . this attribute has been employed to perform all - electron calculations @xcite and to develop a coarse - grained formulation of of - dft for studying crystal defects @xcite . however , higher - order finite - differences shown to be extremely efficient in non - periodic of - dft with the tfw kinetic energy functional @xcite remain unexplored in the context of periodic of - dft simulations , particularly when linear - response kinetic energy functionals like wt and wgc are employed . the electronic ground state in of - dft can be expressed as the solution of a non - linear , constrained minimization problem @xcite . the approaches which have previously been employed to solve this problem include variants of conjugate - gradient @xcite and newton @xcite methods . in these approaches , the techniques used to enforce the constraints include lagrange multipliers @xcite , the penalty method @xcite and the augmented - lagrangian method @xcite . in this work , we present a local real - space formulation and implementation of periodic of - dft . in particular , we develop a fixed - point iteration with respect to the kernel potential for simulations involving linear - response kinetic energy functionals . we develop a parallel implementation of the formulation using higher - order finite - differences . we demonstrate the robustness , efficiency and accuracy of the proposed approach through selected examples , the results of which are compared against existing plane - wave methods . the remainder of this paper is organized as follows . we introduce of - dft in section [ sec : ofdft ] and discuss its real - space formulation in section [ sec : formulation ] . subsequently , we describe the numerical implementation in section [ section : numericalimplementation ] , and validate it through examples in section [ section : examples ] . finally , we conclude in section [ section : conclusions ] . consider a charge neutral system of @xmath0 atoms and @xmath1 electrons in a cuboidal domain @xmath2 under periodic boundary conditions . let @xmath3 denote the positions of the nuclei with charges @xmath4 respectively . the energy of this system as described by of - dft is @xcite @xmath5 where @xmath6 , @xmath7 being the electron density . introducing the parameters @xmath8 and @xmath9 so that different variants of the electronic kinetic energy @xmath10 can be encompassed within a single expression , we can write @xmath11 where @xmath12 is the thomas - fermi energy @xcite , @xmath13 is the von weizsacker @xcite term and @xmath14 is a non - local kernel energy incorporated to make the kinetic energy satisfy the linear - response of a homogeneous electron gas @xcite . they can be represented as @xmath15 where @xmath16 and @xmath17 are parameters , and the constant @xmath18 . on the one hand , the thomas - fermi - von weizsacker ( tfw ) family of functionals with the adjustable parameter @xmath8 is obtained by setting @xmath19 @xcite . on the other hand , kinetic energy functionals which satisfy the lindhard susceptibility function are obtained by setting @xmath20 with appropriate choices of @xmath16 , @xmath17 and the kernel @xmath21 @xcite . in particular , the wang & teter ( wt ) functional @xcite utilizes a density independent kernel , whereas the wang , govind & carter ( wgc ) functional @xcite employs a density dependent kernel . it is common to perform a taylor series expansion of the density dependent kernel @xmath21 about the average electron density @xmath22 @xcite . on doing so , we arrive at @xmath23 where @xmath24 is the order of the expansion , the coefficients @xmath25 and the kernels @xmath26 the second term in eqn . [ eqn : energy : ofdft ] is referred to as the exchange - correlation energy . it is generally modeled in of - dft using the local density approximation ( lda ) @xcite : @xmath27 where @xmath28 is the sum of the exchange and correlation per particle of a uniform electron gas of density @xmath29 . employing the perdew - zunger @xcite parameterization of the correlation energy calculated by ceperley - alder @xcite , the exchange and correlation functionals can be represented as @xmath30 where @xmath31 , and the constants @xmath32 , @xmath33 , @xmath34 , @xmath35 , @xmath36 , @xmath37 and @xmath38 . the final three terms in eqn . [ eqn : energy : ofdft ] represent electrostatic energies @xcite . in periodic systems , they can be expressed as @xmath39 where the summation indices @xmath40 and @xmath41 run over all atoms in @xmath42 and @xmath2 , respectively . the hartree energy @xmath43 is the classical interaction energy of the electron density , @xmath44 is the potential due to the nucleus positioned at @xmath45 , @xmath46 is the interaction energy between the electron density and the nuclei , and @xmath47 is the repulsion energy between the nuclei . the ground state of the system in of - dft is given by the variational problem @xcite @xmath48 where @xmath49 is some appropriate space of periodic functions and @xmath50 represents the constraint on the total number of electrons . the inequality constraint @xmath51 is to ensure that @xmath52 is nodeless , i.e. @xmath52 does not change sign . in this work , we focus on developing a local formulation and higher - order finite - difference implementation for determining the ground - state in periodic of - dft simulations . in this section , we develop a framework for periodic of - dft that is amenable to a linear - scaling real - space implementation . first , we present a local description of the kernel energy and potential in section [ subsec : localreformulationvlr ] . next , we discuss how the electrostatics can be rewritten into local form in section [ subsec : electrostaticreformulation ] . finally , we describe the methodology for determining the of - dft ground - state in section [ subsec : groundstate ] . in simulations where linear - response kinetic energy functionals are employed , the kernel energy @xmath14 as well as the kernel potential @xmath53 \,,\end{aligned}\ ] ] are inherently non - local in real - space . in order to enable a linear - scaling implementation , we start by defining the potentials @xmath54 after approximating the kernels @xmath55 in fourier space using rational functions @xcite , we arrive at @xmath56 where @xmath57 and @xmath58 are solutions of the helmholtz equations @xmath59 under periodic boundary conditions and appropriate choice of complex constants @xmath60 and @xmath61 . above , @xmath62 and @xmath63 thereafter , the kernel potential @xmath64 and the corresponding kernel energy @xmath14 can be calculated in linear - scaling fashion using the expressions @xmath65 \ , , \label{eqn : vlinearresponse : local } \\ t_{lr}(u ) & = & \frac{1}{2}c_f \sum_{m=0}^{l } \sum_{n=0}^{l } \sum_{p=0}^m \sum_{q=0}^n \sum_{r=1}^r c_{mnpq } \int_{\omega } \bigg [ u^{2(m - p+\alpha)}(\bx ) v_{mnq\beta r}(\bx ) + u^{2(n - q+\beta)}(\bx ) v_{mnp\alpha r}(\bx ) \bigg ] \ , \mathrm{d\bx } \nonumber \\ \end{aligned}\ ] ] where @xmath57 and @xmath58 are solutions of the helmholtz equations given in eqns . [ eqn : helmholtz : beta ] and [ eqn : helmholtz : alpha ] , respectively . the electrostatic energies in eqns . [ eqn : eh ] , [ eqn : eext ] and [ eqn : ezz ] are non - local in real - space . moreover , they are individually divergent in periodic systems . to overcome this , we introduce the charge density of the nuclei @xcite : @xmath66 where @xmath67 is the charge density of the @xmath68 nucleus , and the summation index @xmath69 runs over all atoms in @xmath42 . in of - dft calculations , it is common to remove the core electrons and replace the singular coulomb potential with an effective potential @xmath70 , referred to as the pseudopotential approximation @xcite . the absence of orbitals in of - dft requires that the pseudopotential be local , i.e. @xmath70 depends only on the distance from the nucleus . since the pseudopotential replicates the coulomb potential outside the core cutoff radius @xmath71 , @xmath67 has a compact support within a ball of radius @xmath71 centered at @xmath72 @xcite . it follows that @xmath73 using the above definition for the charge densities , we can rewrite the total electrostatic energy as the following variational problem @xmath74 where @xmath75 is the electrostatic potential , @xmath76 is some appropriate space of periodic functions , the second last term accounts for the self energy of the nuclei and the last term corrects for overlapping charge densities . a detailed discussion on the nature of @xmath77 and its evaluation can be found in appendix [ appendix : correct : repulsiveenergy ] . with this reformulation of the total electrostatic energy , we arrive at the variational problem @xmath78 where the functional @xmath79 in the framework described above , the variational problem for determining the ground - state in of - dft can be written as @xmath80 where @xmath81 through this decomposition , the ground - state can be ascertained by solving the electronic structure problem in eqn . [ eqn : groundstateelectronic ] for every configuration of the nuclei encountered during the geometry optimization described by eqn . [ eqn : groundstatesplit ] . below , we discuss the solution strategy for both of these simulation components . consider the variational problem in eqn . [ eqn : groundstateelectronic ] for determining the electronic ground - state . on taking the first variation , we arrive at the euler - lagrange equation @xmath82 where @xmath64 is as given by eqn . [ eqn : vlinearresponse : local ] , and @xmath83 is the lagrange multiplier used to enforce the constraint @xmath84 . further , @xmath75 is the solution of the poisson equation @xmath85 under periodic boundary conditions and @xmath86 the exchange - correlation potential @xmath87 can be decomposed as @xmath88 with @xmath89 and @xmath90 being the exchange and correlation potentials , respectively . even though the notation does not make it explicit , the dependence of @xmath91 on @xmath52 makes eqn . [ eqn : eulerlagrange ] a non - linear problem . it is worth noting that since @xmath92 , the poisson problem defined by eqn . [ eqn : poisson ] with periodic boundary conditions is well - posed . the electronic ground - state can be determined by solving the non - linear eigenvalue problem in eqn . [ eqn : eulerlagrange ] for the eigenfunction corresponding to the lowest eigenvalue . irrespective of the solution technique and choice of kinetic energy functional , @xmath75 needs to be recalculated for every update in @xmath93 . the same is true for @xmath64 when linear - response kinetic energy functionals are employed . therefore , the solution of eqn . [ eqn : eulerlagrange ] requires the repeated solution of the poisson equation in eqn . [ eqn : poisson ] and the complex - valued non - hermitian helmholtz equations in eqns . [ eqn : helmholtz : beta ] and [ eqn : helmholtz : alpha ] . in view of this , the self - consistent field method ( scf ) @xcite commonly utilized in dft calculations is an attractive choice because relatively few iterations are typically required for convergence . however , we have found such an approach to be unstable for both the tfw and wgc kinetic energy functionals , especially as the system size is increased . since the number of helmholtz equations that need to be solved can be significantly large in practice ( e.g. fifty - two in this work ) , they are expected to completely dominate the execution time . in order to mitigate this , we develop a fixed - point method for determining the electronic ground - state when linear response kinetic energy functionals are employed @xcite . this is similar in spirit to the scf method , and is found to converge rapidly , as demonstrated by the examples in section [ section : examples ] . we rewrite the nonlinear eigenvalue problem in eqn . [ eqn : eulerlagrange ] as a fixed - point problem with respect to @xmath64 : @xmath94 \ , , \label{eqn : fixedpoint : map1}\ ] ] where the mappings @xmath95 and @xmath96 & = & c_f \sum_{m=0}^{l } \sum_{n=0}^{l } \sum_{p=0}^m \sum_{q=0}^n \sum_{r=1}^r c_{mnpq } \bigg [ ( m - p+\alpha ) u^{2(m - p+\alpha-1)}(\bx ) v_{mnq\beta r}(\bx ) \ , \nonumber \\ & + & ( n - q+\beta ) u^{2(n - q+\beta-1)}(\bx ) v_{mnp\alpha r}(\bx ) \bigg ] \,.\end{aligned}\ ] ] above , @xmath57 and @xmath58 are solutions to the helmholtz equations given in eqns . [ eqn : helmholtz : beta ] and [ eqn : helmholtz : alpha ] , respectively . the mapping @xmath97 corresponds to the solution of the nonlinear eigenvalue problem in eqn . [ eqn : eulerlagrange ] for a fixed kernel potential @xmath64 . the mapping @xmath98 $ ] corresponds to the calculation of @xmath64 for some given @xmath93 . therefore , the fixed - point of the composite mapping @xmath99 $ ] coincides with the solution of the euler - lagrange equation ( eqn . [ eqn : eulerlagrange ] ) for the electronic ground - state . in order to solve this fixed - point problem , we treat it as a non - linear equation and adopt an iteration of the form @xcite @xmath100 - v_{lr , k } \right)\,,\ ] ] where the index @xmath101 represents the iteration number and @xmath102 is appropriately chosen to ensure / accelerate convergence . once the fixed - point @xmath103 has been determined , @xmath104 can be calculated by solving eqn . [ eqn : mvf ] for @xmath105 . in fig . [ fig : flowchart ] , we present a flowchart that outlines the aforedescribed fixed - point approach . it is worth noting that for the choice of tfw kinetic energy functional ( @xmath19 ) , the solution of eqn . [ eqn : mvf ] coincides with the electronic ground - state . after determining the electronic ground - state , the corresponding energy can be evaluated using the expression @xmath106 \ , \mathrm{d\bx } \nonumber \\ & + & \int_{\omega } \varepsilon_{xc } ( u^(\bx ) ) u^{2}(\bx ) \ , \mathrm{d \bx } + \frac{1}{2 } \int_{\omega}(u^{2}(\bx)+ b(\bx,\br ) ) \phi^{}(\bx ) \ , \mathrm{d\bx } \nonumber \\ & - & \frac{1}{2}\sum_{j } \int_{\omega } b_j(\bx,\br_j ) v_j(\bx,\br_j ) \ , \mathrm{d\bx } + \mathcal{e}_c^(\br ) \,,\end{aligned}\ ] ] where @xmath107 , @xmath108 and @xmath109 are solutions of eqns . [ eqn : helmholtz : beta ] , [ eqn : helmholtz : alpha ] and [ eqn : poisson ] , respectively , for @xmath110 . .,scaledwidth=100.0% ] consider the minimization problem in eqn . [ eqn : groundstatesplit ] for determining the equilibrium configuration of the atoms . during this geometry optimization , the forces on the nuclei can be calculated using the relation @xmath111 where @xmath112 denotes the force on the @xmath68 nucleus and the summation over @xmath113 signifies the @xmath68 atom and its periodic images . additionally , @xmath109 is the solution of the poisson equation in eqn . [ eqn : poisson ] for @xmath114 and @xmath115 corrects for the error in forces due to overlapping charge density of nuclei . the expression for this correction has been derived in appendix [ appendix : correct : repulsiveenergy ] . the second equality in eqn . [ eqn : force : nuclei ] is obtained by using the fact that the energy is stationary with respect to @xmath93 and @xmath75 at the electronic ground - state , and the last equality is obtained by using the spherical symmetry of @xmath116 ( i.e. , @xmath117 ) . since @xmath118 has compact support in a ball of radius @xmath71 centered at @xmath119 , only a finite number of periodic images of the @xmath68 atom have an overlap with @xmath2 . therefore , evaluation of the atomic forces is amenable to a linear - scaling real - space implementation . in this section , we describe a higher - order finite - difference implementation of the formulation presented in the previous section . we restrict our computation to a cuboidal domain @xmath2 of sides @xmath120 , @xmath121 and @xmath122 . we generate a uniform finite - difference grid with spacing @xmath123 such that @xmath124 , @xmath125 and @xmath126 , where @xmath127 , @xmath128 and @xmath129 are natural numbers . we index the grid points by @xmath130 , where @xmath131 , @xmath132 and @xmath133 . we approximate the laplacian of any function @xmath134 at the grid point @xmath130 using higher - order finite - differences @xcite @xmath135 where @xmath136 represents the value of the function @xmath134 at the grid point @xmath130 . the weights @xmath137 are given by @xcite @xmath138 similarly , we approximate the gradient at the grid point @xmath130 using higher - order finite - differences @xmath139 where @xmath140 , @xmath141 and @xmath142 represent unit vectors along the edges of the cuboidal domain @xmath2 . the weights @xmath143 are given by @xcite @xmath144 these finite - difference expressions for the laplacian and gradient represent @xmath145 order accurate approximations , i.e. error is @xmath146 . while performing spatial integrations , we assume that the function @xmath134 is constant in a cube of side @xmath123 around each grid point , i.e. @xmath147 we enforce periodic boundary conditions on @xmath2 by employing the following strategy . in the finite - difference representations of the laplacian and gradient presented in eqns . [ eqn : fd : laplacian ] and [ eqn : gradient : approximate ] respectively , we map any index that does not correspond to a node in the finite - difference grid to its periodic image within @xmath2 . we start with precomputed radially - symmetric and compactly - supported isolated - atom electron densities for each type of atom . we superimpose these isolated - atom electron densities for the initial configuration of the nuclei , and scale the resulting electron density such that the constraint on the total number of electrons is satisfied . we take the pointwise square - root of the electron density so obtained as the starting guess @xmath148 . during the aforedescribed calculation , we only visit atoms whose isolated - atom electron densities have non - zero overlap with @xmath2 . similarly , for every new configuration of atoms encountered during the geometry optimization , we calculate the charge density of the nuclei using the relations @xmath149 where the summation reduces to all atoms whose charge density has non - zero overlap with @xmath2 . the localized nature of the above operations ensures that the evaluation of @xmath148 and @xmath150 scales linearly with the number of atoms . we solve the variational problem in eqn . [ eqn : mvf ] using a conjugate gradient method that was originally developed for dft @xcite and later adopted in simplified form for of - dft @xcite . specifically , we utilize the polak - ribiere update @xcite with brent s method @xcite for the line - search . we refer the reader to appendix [ appendix : nlcgteter ] for further details on the implemented algorithm . for every update in the square - root electron density , we solve the poisson equation in eqn . [ eqn : poisson ] under periodic boundary conditions using the generalized minimal residual ( gmres ) @xcite method with the block - jacobi preconditioner @xcite . since the solution so obtained is accurate to within an indeterminate constant , we enforce the condition @xmath151 for definiteness . in every subsequent poisson equation encountered , we use the previous solution as starting guess . for the complex - valued helmholtz equations in eqns . [ eqn : helmholtz : beta ] and [ eqn : helmholtz : alpha ] , we first separate out each equation into its real and imaginary parts , and then solve the resulting coupled equations simultaneously under periodic boundary conditions using gmres with block - jacobi preconditioners . in every iteration of the fixed - point method , we use the solution of the helmholtz equations from the previous iteration as the starting guess . we accelerate the convergence of the fixed - point iteration by utilizing anderson mixing @xcite , details of which can be found in appendix [ appendix : anderson ] . once the electronic ground - state square - root electron density has been determined , the energy and forces are evaluated using eqns . [ eqn : groundstateenergy ] and [ eqn : force : nuclei ] respectively . while doing so , we restrict the summation over the periodic images to atoms whose charge densities have non - zero overlap with @xmath2 . we solve for the equilibrium configuration of the atoms by using the conjugate gradient method with the polak - ribiere update and secant line search @xcite . we have developed a parallel implementation of the proposed approach using the portable , extensible toolkit for scientific computations ( petsc ) @xcite suite of data structures and routines . within petsc , we have utilized distributed arrays with the star - type stencil option . the communication between the processors is handled via the message passing interface ( mpi ) @xcite . in this section , we validate the proposed formulation and higher - order finite - difference implementation of periodic of - dft through selected examples . henceforth , we shall refer to this framework as rs - fd , which is an acronym for real - space finite - differences . in all the simulations , we employ the goodwin - needs - heine pseudopotential @xcite . in addition , we choose @xmath152 for the tfw functional , and @xmath153 , @xmath154 , @xmath155 , @xmath156 and @xmath157 for the wgc functional . wherever applicable , we compare our results with the plane - wave code profess @xcite . within profess , we utilize a plane - wave energy cutoff of @xmath158 ev , which results in energies and forces that are converged to within @xmath159 ev / atom and @xmath160 ev / bohr respectively . unless specified otherwise , we use sixth - order accurate finite - differences and a mesh size of @xmath161 bohr within rs - fd . we choose a cutoff radius of @xmath162 bohr for the isolated - atom electron densities as well as the charge densities of the nuclei , whereby the enclosed charge for each nucleus is accurate to within @xmath163 . we utilize tolerances of @xmath164 and @xmath165 on the normalized residual as the stopping criterion for the conjugate gradient and gmres methods , respectively . we employ a history of @xmath166 in anderson mixing and a tolerance of @xmath164 on the normalized residual for convergence of the fixed - point method . these parameters and tolerances result in rs - fd energies and forces that are converged to within @xmath167 ev / atom and @xmath167 ev / bohr , respectively . it is worth noting that the aforementioned rs - fd tolerances are highly conservative , i.e. chemical accuracies are achieved even when they are significantly relaxed , as discussed in section [ subsec : scalingperformance ] . we perform all simulations on computer cluster wherein each node has the following configuration : altus 1804i server - 4p interlagos node , quad amd opteron 6276 , 16c , 2.3 ghz , 128 gb , ddr3 - 1333 ecc , 80 gb ssd , mlc , 2.5 " hca , mellanox connectx 2 , 1-port qsfp , qdr , memfree , centos , version 5 , and connected through infiniband cable . we start by verifying convergence of the energy computed by rs - fd with respect to the mesh - size ( @xmath123 ) . as the representative example , we choose a @xmath168-atom face - centered cubic ( fcc ) unit cell of aluminum with lattice constant of @xmath169 bohr , and displace the atom at the corner of the unit cell the origin of the coordinate system to [ @xmath170 @xmath171 @xmath172 bohr . we evaluate the energy of this system as a function of @xmath123 for second and sixth - order accurate finite - difference approximations . in fig . [ fig : convergenceenergy ] , we plot the resulting convergence in energy for the tfw and wgc kinetic energy functionals , with the reference value computed using sixth - order finite - differences and @xmath173 bohr . we observe that sixth - order finite - differences demonstrates significantly higher rates of convergence compared to second - order finite - differences . specifically , the sixth - order scheme obtains convergence rates of @xmath174 and @xmath175 for the tfw and wgc functionals , respectively , whereas the second - order discretization obtains rates of @xmath176 and @xmath177 , respectively . interestingly , the computed convergence rates are not equal to the order of the finite - difference approximation . possible reasons for this include the nonlinearity of the problem , need for finer meshes to obtain the asymptotic convergence rates , the use of trapezoidal rule for integration , and the egg - box " effect . overall , these results indicate that second - order finite - differences are prohibitively expensive for obtaining the chemical accuracies desired in of - dft calculations , thereby motivating higher - order approximations . + next , we verify the convergence of the atomic forces with respect to the mesh - size ( @xmath123 ) . we choose the same example as that used for studying convergence of the energy in section [ subsec : convergenceenergymesh ] . we calculate the force on the displaced atom for the tfw and wgc kinetic energy functionals , and plot the resulting error versus @xmath123 in fig . [ fig : convergenceforce ] . the error is defined to be the maximum difference in the force from that obtained using sixth - order finite - differences with mesh - size of @xmath173 bohr . we again observe that sixth - order finite - differences demonstrates significantly larger convergence rates compared to second - order finite - differences . specifically , the convergence rates obtained by the sixth - order scheme for tfw and wgc are @xmath178 and @xmath179 , respectively , whereas the rates for the second - order approximation are @xmath180 and @xmath181 , respectively . notably , the convergence rates for the force are larger than those obtained for the energy when using a sixth - order discretization . the possible reasons for the convergence rates not matching the finite - difference order are the need for finer meshes for obtaining asymptotic rates , the non - variational nature of the finite - difference approximation , and the egg - box " effect . + overall , we conclude from the results presented in the previous and current subsection that higher - order finite - differences are necessary for performing accurate and efficient electronic structure calculations based on of - dft . indeed , larger convergence rates may be possible as the order of the finite - difference approximation is increased . however , this comes at the price of increased computational cost per iteration due to the reduced locality of the discretized operators and larger inter - processor communication . we have found sixth - order finite - differences to be an efficient choice , which is in agreement with our previous conclusions for the non - periodic tfw setting @xcite . in view of this , we will employ sixth - order finite - differences for all the remaining simulations in this work . we now demonstrate convergence of the fixed - point method for simulations involving the wgc kinetic energy functional . for this study , we choose ( i ) @xmath182-atom system consisting of @xmath183 fcc unit cells of aluminum with lattice constant of @xmath184 bohr ( ii ) @xmath185-atom system consisting of a vacancy in @xmath183 fcc unit cells of aluminum with lattice constant of @xmath184 bohr . for these two examples , we plot in fig . [ fig : andersonmixing : mixparam ] the progression of error during the fixed - point iteration . specifically , we compare the convergence of the basic fixed - point iteration ( i.e. no mixing ) with that accelerated by anderson mixing . within anderson mixing , we choose mixing history size @xmath166 , and mixing parameters @xmath186 and @xmath187 . we observe that anderson mixing significantly accelerates the convergence of the fixed - point iteration , with the mixing parameter @xmath188 demonstrating the best performance . we have found these results to be representative of other calculations utilizing the wgc kinetic energy functional . in view of this , we will utilize anderson mixing with mixing parameter @xmath188 for the remaining simulations in this work . + in fig . [ fig : andersonmixing : mixhist ] , we compare the convergence of the anderson accelerated fixed - point iteration for mixing histories of different sizes . specifically , we choose @xmath166 , @xmath189 , and @xmath190 for this study . we observe that the size of the mixing history does not have any noticeable impact on the fixed - point iteration . in fact , the plots of the error versus iteration number in fig . [ fig : andersonmixing : mixhist ] are nearly identical . overall , we conclude that the fixed - point iteration accelerated with anderson mixing is extremely robust and efficient . in particular , the error decreases rapidly , and approximately @xmath191 iterations are sufficient to obtain the desired chemical accuracy in energies and forces . indeed , the energies are converged to within @xmath192 ev / atom and the forces are converged to within @xmath193 ev / bohr for a fixed - point iteration error of @xmath194 in figs . [ fig : andersonmixing : mixparam ] and [ fig : andersonmixing : mixhist ] . + in this work , we have proposed a fixed - point problem with respect to @xmath64 for simulations involving linear - response kinetic energy functionals . however , it is also possible to develop an analogous fixed - point problem with respect to @xmath93 . in table [ table : mixcomparison ] , we compare the performance of the fixed - point iterations with respect to @xmath93 and @xmath64 for the wgc functional . in both cases , we accelerate the iteration using anderson mixing with @xmath166 . it is clear that the relative performance of the two fixed - point iterations is system dependent . however , we have found that the iteration with respect to @xmath64 is significantly more robust than the one with @xmath93 . therefore , we employ the fixed - point iteration with respect to @xmath64 for determining the electronic ground - state in simulations involving linear - response kinetic energy functionals . .number of steps required to reduce the error to @xmath164 in the fixed - point iterations with respect to @xmath52 and @xmath195 . anderson mixing with @xmath166 has been employed in both cases . [ cols="^,^,^",options="header " , ] in the local reformulation of the electrostatics presented in section [ subsec : electrostaticreformulation ] , the repulsive energy can be expressed as @xmath196 where the second term accounts for the self energy of the nuclei . using eqn . [ eqn : b : pseudopotential ] , we arrive at @xmath197 above , the summations indices @xmath40 and @xmath69 run over all atoms in @xmath42 . if the charge density of the nuclei do not overlap , eqn . [ eqn : b : ezz ] can be rewritten as @xmath198 which is exactly the expression given in eqn . [ eqn : ezz ] for the repulsive energy prior to reformulation . however , the use of relatively ` soft ' pseudopotentials which are attractive because of the significant reduction in the number of basis functions required for convergence can frequently result in overlapping charge density of the nuclei . even in this situation , the repulsive energy in ab - initio calculations is calculated by treating the nuclei as point charges ( i.e. , eqn . [ eqn : repulsiveenergyappendix ] ) . since the electrostatic reformulation in this work does make this distinction between overlapping and non - overlapping charge density of the nuclei , we present a technique below that reestablishes agrement . we start by generating a ` reference ' charge density @xmath199 which is the superposition of spherically symmetric and compactly supported ` reference ' charge densities @xmath200 . these nuclei - centered charge densities satisfy the relations @xmath201 thereafter , the correction to the repulsive energy can be expressed as @xmath202 a direct computation of this energy correction will scale quadratically with respect to the number of atoms . in order to enable linear - scaling , we rewrite eqn . [ eqn : repulsivecorrection ] as @xmath203 where @xmath204 is the solution to the poisson equation @xmath205 with periodic boundary conditions . the potential @xmath204 so calculated is accurate to within a constant , which can be determined by evaluating @xmath206 at any point in space . the correction to the forces on the nuclei @xmath207 can be represented as @xmath208 \,\mathrm{d\bx } \nonumber \\ & = & \frac{1}{2 } \sum_{j ' } \int_{\omega } \bigg [ \nabla \tilde{b}_{j'}(\bx,\br_{j ' } ) \left(v_c(\bx,\br)- \tilde{v}_{j'}(\bx,\br_{j'})\right ) + \nabla b_{j'}(\bx,\br_{j ' } ) \left(v_c(\bx,\br)+v_{j'}(\bx,\br_{j'})\right ) \nonumber \\ & & + \nabla v_{c , j'}(\bx,\br_{j ' } ) \left(\tilde{b}(\bx,\br)+b(\bx,\br)\right ) + b_{j'}(\bx,\br_{j ' } ) \nabla v_{j'}(\bx,\br_{j ' } ) - \tilde{b}_{j'}(\bx,\br_{j ' } ) \nabla \tilde{v}_{j'}(\bx,\br_{j ' } ) \bigg ] \,\mathrm{d\bx } \nonumber \,,\end{aligned}\ ] ] where the summation @xmath113 is over @xmath68 atom and its periodic images . additionally , @xmath209 it is important to note that even with these corrections to the energy and forces , the overall of - dft formulation maintains its linear - scaling nature with respect to the number of atoms . in algorithm [ algo : nlcg ] , we present the conjugate gradient method implemented in rs - fd to solve the variational problem in eqn . [ eqn : mvf ] . this differs from the standard non - linear conjugate gradient method @xcite in that it is able to handle the constraints @xmath84 and @xmath51 . [ algo : nlcg ] input * : @xmath210 , @xmath211 , @xmath195 and @xmath212 + @xmath213 + ( ) @xmath214 @xmath215 , where @xmath216 denotes the inner product + @xmath217 + @xmath218 + @xmath219 + @xmath220 + @xmath221 + * output * : @xmath222 the fixed - point problem in eqn . [ eqn : fixedpoint : map1 ] can be rewritten as @xmath223 - v_{lr } \,.\ ] ] this equation can be solved using an iteration of the form @xcite @xmath224 where @xmath102 is chosen to approximate the inverse jacobian . in multi - secant type methods , @xmath102 is set to the solution of the constrained minimization problem @xcite @xmath225 where @xmath226 \ , , \nonumber \\ y_k & = & [ f(v_{lr , k - m+1 } ) - f(v_{lr , k - m } ) , \ldots , f(v_{lr , k } ) - f(v_{lr , k-1 } ) ] \ , . \nonumber \end{aligned}\ ] ] in the above equations , @xmath227 represents the mixing history . the solution of this variational problem is @xmath228 in the specific case of anderson mixing @xcite , @xmath229 is set to @xmath230 , where @xmath40 is a @xmath231 identity matrix . this leads to the update formula : @xmath232 65 natexlab#1#1[2]#2 , , ( ) . , , ( ) . , , , , , . ( ) . , , ( ) . , , , ( ) , , , ( ) . , , in : ( ed . ) , , volume of _ _ , , , pp . . , ( ) . , , , , ( ) . , , , ( ) . , , ( ) . , , , , ( ) . , , , ( ) . , , , , ( ) . , , ( ) . , , , ( ) . , , , ( ) . , , , , , , ( ) . , , ( ) . , , , , ( ) . , , , , ( ) . , , , ( ) . , ( ) . , , , ( ) . , , ( ) . , , , ( ) . , , ( ) . , , , ( ) . , , ( ) . , , , ( ) . , ( ) . , , ( ) . , , , , , . , ( ) . , , , ( ) . , , , , , ( ) . , , ( ) . , , , volume , , . , ( ) . , , , , , , , , , , , , , argonne national laboratory , . , , , , in : , , ( eds . ) , , , , pp . . , , , , volume , , . , , , ( ) . , , ( ) . , , , . , ( ) . , , volume , , .	we present a real - space formulation and higher - order finite - difference implementation of periodic orbital - free density functional theory ( of - dft ) . specifically , utilizing a local reformulation of the electrostatic and kernel terms , we develop a generalized framework for performing of - dft simulations with different variants of the electronic kinetic energy . in particular , we propose a self - consistent field ( scf ) type fixed - point method for calculations involving linear - response kinetic energy functionals . in this framework , evaluation of both the electronic ground - state as well as forces on the nuclei are amenable to computations that scale linearly with the number of atoms . we develop a parallel implementation of this formulation using the finite - difference discretization . we demonstrate that higher - order finite - differences can achieve relatively large convergence rates with respect to mesh - size in both the energies and forces . additionally , we establish that the fixed - point iteration converges rapidly , and that it can be further accelerated using extrapolation techniques like anderson s mixing . we validate the accuracy of the results by comparing the energies and forces with plane - wave methods for selected examples , including the vacancy formation energy in aluminum . overall , the suitability of the proposed formulation for scalable high performance computing makes it an attractive choice for large - scale of - dft calculations consisting of thousands of atoms . finite - differences , real - space , fixed - point , anderson mixing , conjugate gradient , electronic structure .
this is an exciting time for the maturing field of gravitational wave ( gw ) physics . the network of ground - based laser interferometer gw detectors ( gwd ) is making rapid progress towards its goal of producing advanced instruments sensitive enough to make monthly gw detections in the 10hz to 10khz frequency range by the end of the decade @xcite . simultaneously , there has recently been a claim by the bicep2 experiment that the measurement of the b - modes in the cosmic microwave background polarization can be the signature of the primordial gws produced by inflation @xcite . if confirmed , this would add to the original evidence for gws from 1975 by hulse and taylor who observed that the rate of change of the orbital period of a binary star system precisely agrees with the predictions of gr @xcite . moreover , the elisa space - based gravitational wave detector @xcite together with the promising pulsar timing technique @xcite will be very important in the coming decades for the investigation of very massive objects and other gw sources in the milli- to micro - hz frequency ranges , respectively . gws are dynamic strains in space - time that travel at the speed of light and are generated by non - axisymmetric acceleration of mass . they perturb the flat minkowski metric describing space - time . the effect is the production of a dimensionless strain between two inertial masses located at a proper distance @xmath0 from one another so that their distance changes as : @xmath1 in the late 20th century following the era of the operation of bar detectors for gw detection @xcite , large laser interferometers were identified as the promising route forward because of the very high strain sensitivities that could be achieved over a wide frequency band . r. weiss produced the first detailed design study in 1972 of a large scale laser interferometer for gw detection , complete with calculations of fundamental noise sources @xcite . then , following the work of forward @xcite who built the first laser interferometric gw detector prototype , many groups around the world proceeded to study the benefits of laser interferometry , build new prototypes , perfect the design , and push technology development @xcite . the ligo detectors in the u.s . @xcite , the virgo detector in italy @xcite , the geo600 detector in germany @xcite and the tama300 @xcite detector in japan formed the first generation laser - interferometric gwd network . construction of these projects began in the mid-1990s and then progressed in sequential commissioning and data - taking phases in the 2000s . these first - generation detectors achieved their original instrument sensitivity goals and are now undergoing major hardware upgrades to expand thousand - fold the observable volume of the cosmos . the strain sensitivities achieved by these initial detectors is shown in figure [ fig : sensitivity ] . this paper describes the common design basics of the detectors and provides an overall summary of the current status of the worldwide network of gwds , including a timeline for operations . the next four proceedings focus on the particular features and individual status of each of the advanced versions of these detectors @xcite . a direct detection of gravitational waves themselves has not yet been made , but this is also not a surprise . the rate estimates for coalescing binary neutron stars , for instance , predict a detection probability of one event in 100 years for the initial detector sensitivities @xcite . about two years worth of double - coincidence data were collected . nonetheless , the ligo and virgo collaborations have already produced astrophysical results from the data collected thus far . they have placed an upper limit that beats previous best estimates of the fraction of spin - down power emitted in gws from the crab pulsar @xcite . they have also placed an upper limit at 100hz on the normalized energy density of the stochastic gw background of cosmological and astrophysical origin , a result otherwise inaccessible to standard observational techniques @xcite . the last decade brought great advances in demonstrating the experimental feasibility of achieving the strain sensitivities required to witness astrophysical events and informed the design of today s generation of gwds . both ligo and virgo are re - using the infrastructure from the initial generation detectors and are replacing the hardware within the vacuum system . the upgraded detectors are called advanced ligo @xcite and advanced virgo @xcite , respectively . tama300 as well as clio , a prototype cryogenic laser interferometer @xcite , informed the design of a brand new detector called kagra , the first underground , cryogenic laser interferometer @xcite . geo600 is keeping its infrastructure and most of its initial generation hardware , but is carrying out upgrades to demonstrate advanced techniques in a program called geo - hf @xcite . in addition , a proposal to expand ligo s baseline by building an interferometer in india is moving forward . the typical advanced detector configuration is that of a dual - recycled fabry - perot michelson ( drfpm ) laser interferometer as depicted in figure [ fig : ifoschematic ] . a power amplified , and intensity and frequency stabilized nd : yag solid state laser system injects linear - polarized 1064 nm light into a triangular input mode - cleaner ( imc ) cavity . the imc suppresses laser frequency noise and provides spatial filtering of the laser beam to reduce beam jitter that could otherwise couple to the gw readout . a beam splitter ( bs ) sends the beam to the two fabry - perot arms , which are made of an input test mass mirror ( itm ) and an end test mass mirror ( etm ) . both arms are of km - scale lengths and are set to maintain nearly perfect destructive interference of the recombined light at the anti - symmetric ( as ) port which carries the gw information . here , the beam is directed to an output mode - cleaning ( omc ) cavity and then onto a photo - detector . the omc transmits only the signal - carrying light to improve the signal - to - noise ratio . a power recycling mirror at the symmetric port directs the constructively - interfered light back into the interferometer . the transmissivity of the power recycling mirror is set to match the losses of the main optics to create a nearly critically coupled cavity . a signal recycling mirror at the anti - symmetric port creates an additional cavity which can be used to adjust the storage time of the gravitational wave signal in the interferometer and thus the frequency response of the detector . the signal recycling mirror has a transmissivity selected to compromise between high and low frequency sensitivity based on thermal noise and laser power . until a few years ago all detectors used heterodyne readout . since approximately 2008 as part of intermediary upgrades to the initial detectors @xcite , homodyne ( _ dc _ ) readout was implemented together with the addition of an omc @xcite . nearly the entire interferometer is enclosed in an ultra high vacuum ( uhv ) system to render phase noise of residual gas unimportant and to keep the optics free of dust . the primary interferometer optics are suspended as pendulums to decouple them from ground motion so that they act like free masses in the horizontal plane at the frequencies in the gw detection band . to minimize the impact of thermal noise , the mirror suspensions are designed to minimize dissipation and , in the case of kagra , operated at cryogenic temperatures . each mirror is equipped with actuators for coarse and fine control of the mirror position and orientation . a feedback control system is implemented to hold the system sufficiently near the intended operating point so that the response to residual deviations remains linear . calibration of the detector must take into account the action of the control system and the frequency response of the detector @xcite . the various length ( illustrated in figure [ fig : ifoschematic ] ) and angular degrees of freedom are sensed through the use of radio - frequency sidebands on the carrier light that are created through phase modulation by electro - optic modulators . the differential arm length signal is sensitive to gravitational waves , and is sensed using homodyne readout in transmission of the omc , where the gw signal is encoded as power variations of the light . the standard technique for locking optical cavities is the pound hall method of laser frequency stabilization @xcite . although the interferometer is an analog instrument , it is interfaced through a digital control system which allows complex filters to be implemented and tuned from the control room . the sources of noise that contaminate the detector s output can be grouped into two categories : displacement noise and sensing noise . displacement noises are those that create real motion of the mirrors , while sensing noises are those that arise in the process of measuring the mirrors position . the primary displacement noise that plagues terrestrial laser interferometers at very low frequencies is motion of the ground , i.e. seismic noise . thermal motion of the mirrors , their dielectric coatings and suspensions as well as quantum radiation pressure noise are the other two types of displacement noise which dominate in the low- to mid - frequency range @xcite . the primary sensing noise is shot noise that arises from the poisson statistics of photon arrival at the photodetector . figure [ fig : designcurves ] shows the spectra of these ultimate limits to the performance of each of the gw detectors in the advanced detector network . each curve reflects the incoherent sum of fundamental noise sources , which gives a likely best limit to performance . the actual sensitivity will depend also on technical noise sources . the narrow lines in each of the noise curves represent the thermally - excited violin modes of the test mass suspension fibers . quantum noise and thermal noise play the dominant roles in limiting the sensitivity of each of the detectors . the frequency at which there is a dip in the noise together with the shapes of the noise curves are largely affected by the signal recycling parameters , which may be adjusted during the course of operation of the advanced detectors . we postpone the detailed description of each noise curve to the proceedings dedicated to each detector . table [ tab : overview ] provides an overview comparing some of the major properties of each of the detector designs . [ cols="<,<,<,<,<",options="header " , ] [ tab : overview ] the advanced detectors form a four - site network which is crucial for gw signal characterization . the sky coverage depends on the detector locations and orientations . the two ligo detectors are nearly aligned for maximum correlation , but they are relatively close together which results in signals that have largely redundant information about the source direction and character . for this reason , only the four - site network can provide full sky coverage @xcite . another important feature of a detector network is sky localization for electro - magetic follow ups and multi - messenger investigations . in the case of signals from coalescing binary neutron star systems , the two advanced ligo detectors , advanced virgo , and kagra would provide a localization accuracy of 10 @xmath2 for @xmath3 of the sources , while a five site network , including ligo india , would improve the accuracy to 5 @xmath2 for @xmath4 of the sources @xcite . this network of gw detectors together with the joint detection of the electro - magnetic counterparts is poised to open a promising window to gravitational wave astronomy . figure [ fig : timeline ] shows a timeline of the construction , commissioning , and science run stages of each of the gwds . the down times for each of the detectors is a bit staggered , with geo600 serving the role of being the sole detector online for the entirety of the period when the other detectors are offline . commissioning and science run stages of the advanced detectors will be interspersed to allow for possible early results once astrophysically interesting sensitivities are reached , but before design sensitivity is reached @xcite . the early science runs are likely to start in late 2015 with the advanced ligo detectors , and advanced virgo and kagra will join in due time . together with the promise of continuing to add to the collection of upper limits on gw emission in the era of advanced detectors , the first direct detection of gws is expected to be made in only a couple years time from now . a first detection is expected to witness an event such as a binary neutron star coalescence @xcite . the authors gratefully acknowledge the support of the united states national science foundation for the construction and operation of the ligo laboratory ; the science and technology facilities council of the united kingdom , the max - planck - society , and the state of niedersachsen / germany for support of the construction and operation of the geo600 detector ; the italian istituto nazionale di fisica nucleare and the french centre national de la recherche scientifique for the construction and operation of the virgo detector ; and the japan society for the promotion of science ( jsps ) core - to - core program a , advanced research networks , and grant - in - aid for specially promoted research of the kagra project . this document has been assigned ligo document number ligo - p1400153 . 10 url # 1#1urlprefix[2][]#2 abadie j _ et al . _ 2010 _ classical and quantum gravity _ * 27 * 173001 + geo - hf logbook , p.60 https://intranet.aei.uni - hannover.de / geo600/geohflogbook.nsf/% f5b2cbf2a827c0198525624b00057d30/4837a612ac990060c12575ce004e70fd[https://intranet.aei.uni - hannover.de / geo600/geohflogbook.nsf/% f5b2cbf2a827c0198525624b00057d30/4837a612ac990060c12575ce004e70fd ] the ligo scientific collaboration and the virgo collaboration 2013 prospects for localization of gravitational wave transients by the advanced ligo and advanced virgo observatories http://arxiv.org/abs/1304.0670	ground - based laser interferometers for gravitational - wave ( gw ) detection were first constructed starting 20 years ago and as of 2010 collection of several years worth of science data at initial design sensitivities was completed . upgrades to the initial detectors together with construction of brand new detectors are ongoing and feature advanced technologies to improve the sensitivity to gws . this conference proceeding provides an overview of the common design features of ground - based laser interferometric gw detectors and establishes the context for the status updates of each of the four gravitational - wave detectors around the world : advanced ligo , advanced virgo , geo600 and kagra .
iras 16293@xmath12422 is one of the best studied low - mass young stellar objects ( ysos ) @xcite . it is deeply embedded in the ldn 1689n cloud in ophiuchus ( distance 160 pc ) and classified as an extreme class i / class 0 yso @xcite . it is a low - luminosity ( 27 @xmath23 , @xcite ) binary system ( hereafter , i16293a and i16293b ) with a separation of 840 au ( @xmath24 ) and individual stellar masses of 0.5 m@xmath13 @xcite . two powerful outflows emanate from the system . at least three physically and chemically different regions can be recognised in a @xmath25 ( 3000 au ) beam @xcite : ( i ) a cold , relatively low - density outer envelope ( @xmath261020 k , @xmath27(h@xmath28@xmath29 @xmath30 ) which marks the transition into the extended parental cloud ldn 1689n ; ( ii ) a warmer circumbinary envelope of size @xmath31 au ( @xmath32 k , @xmath27(h@xmath33@xmath34 @xmath30 ) ; and ( iii ) a warm , dense core of about 5001500 au ( @xmath35 k , @xmath27(h@xmath36 @xmath30 ) which traces the interaction of the outflow and the stellar radiation with the inner part of the circumbinary envelope . this picture emerged from excitation analysis of species tracing these distinct regions and conditions , assuming uniform temperatures and densities for each . continuous descriptions of the density and temperature as function of radius in the collapsing envelope were developed by @xcite and @xcite on scales from 30 to 5000 au . in this paper , we present new observations and modeling , that allows us to independently derive the density and temperature structure in the extended envelope around the ( pre- ) protostellar cores through continuum photometry and imaging of the dust , and heterodyne spectroscopy of the molecular gas . the resulting density and temperature distribution ranges from the warm inner envelope region ( 100 au ) to the coldest ( 12 k ) outer envelope regions ( 7300 au ) and serves as a basis for excitation and abundance analysis of the line data . another component of the iras 16293@xmath12422 system is located 80@xmath37 east and 60@xmath37 south of i16293a . this small clump , i16293e , has strong dco@xmath2 , nh@xmath38 , and nh@xmath4d emission @xcite connected to the i16293a core , and significant submillimeter continuum emission from dust @xcite . single - dish observations of nh@xmath38 in a @xmath39 beam indicate a temperature @xmath40 k @xcite . millimeter - interferometric data , sampling smaller scales and less sensitive to the warmer extended envelope , yield @xmath41 k and @xmath42 @xmath30 if homogeneous conditions are assumed @xcite . because no far - infrared ( fir ) or submillimeter point source has been found embedded in i16293e , it is classified as a pre-(proto)stellar core . in this paper , we present submillimeter images that constrain the density structure of i16293e . it is difficult to study chemical evolution in the early protostellar stages , because during collpase the temperature stays low at @xmath43 k while the density increases to @xmath44h@xmath45 @xmath30 . under these conditions the line emission of many molecules is dominated by a warmer outer envelope around the prestellar core . also , many molecules in the cold core will condense on dust grains . eventually , only h@xmath4 , h@xmath46 , and their isotopomers hd and h@xmath4d@xmath2 will remain in the gas phase . this leads to a significant enhancement of h@xmath4d@xmath2 through the deuterium exchange reaction @xmath47 since the backward reaction becomes negligible at low temperatures @xcite . h@xmath4d@xmath2 is thought to play a pivotal role in the deuteration of molecules @xcite and the detection of h@xmath4d@xmath2 in the class 0 yso ngc 1333 iras 4a @xcite is an important confirmation of the cold gas - phase chemical networks . the h@xmath4d@xmath2 enhancement is reflected in the high abundance ratios that are generally observed of , e.g. , [ dco@xmath2]/[hco@xmath2 ] , [ nh@xmath4d]/[nh@xmath38 ] , [ dcn]/[hcn ] , and [ n@xmath4d@xmath2]/[n@xmath4h@xmath2 ] in cold prestellar cores and class 0 ysos ( e.g. , @xcite ) . models predict that in the collapse stage the abundances of many deuterated radicals and molecules increase sharply after their non - deuterated versions start to get heavily depleted on dust grains , then reach a peak and start to decrease as the deuterated molecules too condense on the dust ( e.g. , @xcite ) . reactions with atomic deuterium on the grain surface may further enhance the abundance of deuterated species in the ice @xcite . after the formation of a yso , the condensed molecules may evaporate from grains , resulting in a temporary increase in the gas - phase deuteration fraction . the deuteration may decline after @xmath50 years @xcite when higher temperature chemistry becomes effective . in addition , in such hot ( @xmath51200 k ) regions the reversal of reaction ( [ e1 ] ) becomes dominant and very little new fractionation is expected to occur . we report in this paper observations of several key deuterated species in i16293a and i16293e . water is an important species in the early stages of star formation . h@xmath4o is the most important hydride molecule in oxygen chemistry and models predict that h@xmath4o plays a dominant role in the energy balance and chemical evolution during star formation . water will have the largest abundance and excitation contrast between the protostellar source and the surrounding cloud ( e.g. , @xcite ) . measurements of the rotational and ro - vibrational transitions of h@xmath4o in star - forming regions have recently become available from the infared space observatory ( iso ; e.g. , @xcite ) and the submillimeter wave astronomical satellite ( swas ; e.g. , @xcite ) . in particular the heterodyne velocity - resolved measurements of the ground - state transition of ortho - h@xmath4o with swas allow a direct determination of the water abundance throughout the envelope . the hdo @xmath5@xmath6 ground - state transition at 465 ghz ( @xmath52 k , @xmath53 @xmath30 for @xmath54 k , @xcite ) is an excellent tracer of high density gas at low temperatures where the abundance of hdo relative to water is enhanced . ground - based observations of the hdo ground - state transition are still very sparse . to date , this transition has only been observed in the high - mass star - forming regions orion - kl @xcite , w3(oh ) , and w3(h@xmath4o ) @xcite . the latter study indicates that the [ hdo]/[h@xmath4o ] abundance ratio is comparable to that found in hot cores and is not a sensitive indicator of the evolutionary stage in high - mass star formation . this could mean either that the w3 cloud always stayed warm or that the [ hdo]/[h@xmath4o ] ratio retrusn to thermal equilibrium faster than ratios like [ dcn]/[hcn ] @xcite . for a low - mass class 0 yso like i16293a the situation may be different , because the time scale to reach steady - state may be much longer . in this paper , we present observations of hdo and h@xmath4o of i16293a and e. in this paper we present a detailed study of the physical and chemical structure of the low - mass star - forming cloud ldn 1689n with particular emphasis on the h@xmath4o / hdo chemistry and deuterium fractionation . we present extensive submillimeter continuum photometry and maps revealing the warm and cold dust around i16293a , b and i16293e ( [ obs_cont ] ) . we report the detection of the ground - state lines of ( deuterated ) water and a tentative detection of h@xmath4d@xmath2 toward i16293a as well as weak hdo emission in the prestellar core i16293e ( [ obs_line ] ) . these data are combined with pointed observations of co and hco@xmath2 isotopomers to determine the temperature and density structure throughout the envelope ( [ results ] ) . we address the two outflows and their origin in [ co_outflows ] . our deep observations of the hdo and h@xmath4o ground - state transitions are used to study the [ hdo]/[h@xmath4o ] ratio throughout the envelope ( [ hdo_h2o ] ) . a simple chemistry network is used to model the deuterium chemistry and in particular the water fractionation ( [ simple_chem_model ] ) . finally , we discuss the evolutionary difference between i16293a , i16293b , and i16293e ( [ nature_of ] ) , and summarize the conclusions in [ conclusion ] . all continuum observations were made with the james clerk maxwell telescope ( jcmt ) , using the single channel bolometer ukt14 and the versatile submillimetre common - user bolometer array ( scuba ) under good weather conditions . the ukt14 measurements were done during the winters of 19911993 . this instrument was for a long time the common - user bolometer system and is described in @xcite . it was replaced in 1997 by scuba which was designed both for photometry and fast mapping @xcite . the photometry of i16293e was done with ukt14 during spring 1993 . the chop throw was 90@xmath37 in declination and the calibrators were uranus and the nearby secondary calibration standard i16293a @xcite . the results are given in table [ t1 ] . ukt14 maps were obtained at 1.1 mm , 800 , 750 , 450 , and 350during several observing runs in 1991 and 1992 . most of these maps have poor sensivitity and were mainly used to derive the submillimeter position of i16293a . however , even with the somewhat limited signal - to - noise ( s / n ) these provide a better estimate of the background subtracted flux of i@xmath55a than single position photometry . at 800 we attempted to go deep and completely map the dust cloud surrounding the bright submillimeter source . because the dust emission is very extended ( see the 850 @xmath0 m scuba map in fig . [ f1 ] ) we combined six maps of the region , centered on i16293a or i16293e . each map was bound by pointing observations of the nearby blazars 1514@xmath1241 and 1730@xmath1130 . typical pointing drifts were less than 2@xmath37 . these errors were removed in the data reduction process assuming that the pointing drifts linearly in azimuth and elevation as a function of time . all maps were made on - the - fly , scanned in azimuth with a cell size of 4@xmath37 and a chop throw of 40@xmath37 in the scan direction . with a single pixel instrument it is difficult to map such an extended region . the final map was therefore reduced with the dual beam maximum entropy ( dbmem ) algorithm written by john richer @xcite for jcmt . the dbmem map was baseline subtracted and recalibrated by using the integrated intensity in a @xmath56 area around i16293a from maps reduced with nod2 @xcite . the 800-map is shown in fig . [ f1 ] ( top panel ) . from comparison with scuba maps ( see below ) , it is clear that the 800-map has failed to recover the faint , relatively uniform , low - level emission surrounding both cloud cores . this does not affect the morphology of the map , but does have a large effect on integrated intensities of the surrounding ldn 1689n cloud . the low - level emission is missing because the 800-maps were taken with the same rather short chop throw , and the individual maps have a noise level comparable or higher to that of the low - level extended emission . it is also well known that the nod2-algorithm is not very sensitive to relatively uniform extended emission on spatial scales of several times the chop throw @xcite . additional maps were obtained with scuba during spring 1997 as part of commissioning jiggle- and scan - map observing modes . we have complemented these maps with additional maps obtained from the jcmt archive in the time period 19971998 . all maps were taken under excellent sky conditions ( atmospheric opacities @xmath57 ) , and all the maps in the archive were taken with a 120@xmath37chop throw , typically in azimuth . most of the maps of i16293e were also obtained with a 120@xmath37 chop . the pointing was checked before and after each map using the same blazars as for our ukt14 observations . some of the archive maps , however , were taken without pointing observations . these were aligned with the pointing corrected average before adding them into the final data set . calibration and beam characterization is based on beam maps of uranus obtained with the same chop throw . the uranus maps give a half power beam width of 14.5@xmath37 and 7.8@xmath37 for 850 @xmath0 m and 450 @xmath0 m , respectively . both the i16293a , b and i16293e cloud cores are rather extended and no map is completely free of emission in the off positions . any bolometer that showed evidence for or was suspected of being contaminated by the chop was therefore blanked in the data reduction . although this was done as carefully as possible , it is clear that the outskirts of our final mosaiced image ( fig . [ f1 ] ) may still be affected , especially since the dust emission extends beyond the area we mapped to the north and northeast of both i16293a and i16293e . the cloud core surrounding i16293e also extends further east than is covered by our maps . the basic data reduction was done using the scuba software reduction package surf @xcite as explained in the scuba map reduction cookbook @xcite . in total we used 16 data sets for the 850 @xmath0 m map and 13 for the 450 @xmath0 m map . in the final coadd we adjust the pointing in each map ( shift and add ) to ensure that the final map is not broadened by small pointing errors . the rms noise levels in the final maps are difficult to estimate , because there are no emission - free regions in the maps , but we estimate them to be below 20 mjy beam@xmath10 at 850 @xmath0 m and below 80 mjy beam@xmath10 at 450 @xmath0 m . the final maps were converted to fits and exported to miriad @xcite for further analysis . in order to correct for the error beam contribution , especially at 450 @xmath0 m , we deconvolved the maps using clean and spherically symmetric model beams derived from beam maps of uranus observed in stable night time conditions . at 850 @xmath0 m we use hpbws of 14.5@xmath37 , 55@xmath37 , and 150@xmath37 , with amplitudes of 0.985 , 0.014 and 0.001 , respectively . at 450 @xmath0 m the hpbws are 7.8@xmath37 , 34@xmath37 , and 140@xmath37 , with amplitudes of 0.964 , 0.029 , and 0.007 , respectively . the final maps were restored with a 14@xmath37beam at 850 and an 8@xmath37 beam at 450 @xmath0 m . the peak flux densities in the restored maps are 16.0 jy beam@xmath10 and 76.1 jy beam@xmath10 toward i16293a at 850 and 450 @xmath0 m , respectively , while the peak fluxes of i16293e are 1.4 jy beam@xmath10 and 4.35 jy beam@xmath10 at 850 and 450 @xmath0 m , respectively . spectra of co @xmath15=21 ( 230.538000 ghz ) and isotopomers , as well as the @xmath15=32 transitions of dco@xmath2 ( 216.112605 ghz ) , hco@xmath2 ( 267.557619 ghz ) , and h@xmath3co@xmath2 ( 260.255478 ghz ) toward i16293a and i16293e were retrieved and reduced from a search of all released data in the jcmt archive at the canadian astronomy data center taken with the sis receiver rxa2 and its successor rxa3 . deep observations of the hdo @xmath58@xmath59 ( 225.896720 ghz ) and @xmath60@xmath59 ( 241.561550 ghz ) lines were made with the receiver rxa3 in 2001 april . we also acquired a large , fully sampled map in the co 21 line which is centered on i16293e and covers the outflows from i16293a and b. the map was made on - the - fly with a bandwidth of 125 mhz , a sampling of 5 in the scan direction , and a cross - scan step size of 10 . the final map was made from a series of sub - maps scanned either in ra or dec . the integration time per point was 5 seconds , but in poor weather conditions we often made multiple coverages . all observations were done in position switch mode with the reference position 600@xmath37 or 800@xmath37 west of the submillimeter position of i16293e . at the same time we have also obtained spectra toward i16293a , which is a pointing and spectral line standard . the relative calibration of the spectra of the co 21 map is accurate within 5% , because we overlapped the sub - maps . in 1999 april we obtained additional deep spectra of co and @xmath3co 21 ( 220.398677 ghz ) at the peak positions in the outflow lobes in order to get more accurate estimates of the @xmath61co optical depth in the high velocity gas . most of the spectra were obtained with a bandwidth of 125 mhz , corresponding to a velocity resolution of 0.1 km s@xmath10 . the spectra are calibrated in @xmath62 ( i16293a ) or @xmath63 ( i16293e ) and shown in figs . [ f2 ] and [ f3 ] . the results of gaussian fits to the line profiles are summarised in table [ t2 ] . for some of the lines with symmetric profiles ( e.g. , hdo ) the emission and the absorption components were fitted separately . for most of the complex line profiles we only list the integrated emission and the velocity of the maximum self - absorption . note that the phase - lock instability of rxa2 and rxa3 may cause an instrumental broadening up to 0.5 km s@xmath10 . the n@xmath4h@xmath2 @xmath15=43 ( 372.672509 ghz ) , h@xmath4d@xmath2 @xmath7@xmath8 ( 372.421340 ghz ) , dco@xmath2 54 ( 360.169881 ghz ) , hco@xmath2 43 ( 356.734288 ghz ) , and h@xmath3co@xmath2 43 ( 346.998540 ghz ) rotational transitions were observed in 2000 april with the jcmt using the dual channel sis receiver rxb3 . the receiver was operated in single side band ( ssb ) mode where the image side band is rejected and terminates in a cold load . the main beam efficiency at these frequencies is about 60% , while the beam width is 1413@xmath37for 345372 ghz . the lines were observed in beamswitching mode switching over 180@xmath37 . the n@xmath4h@xmath2 and h@xmath4d@xmath2 lines were observed simultaneously with relatively good atmospheric transparency ( @xmath64 ) , but with only 30 minutes of useful integration time above an elevation of 40 degrees . rxb3 was changed to tunerless mixers in 1999 and the performance near the upper band edge has decreased dramatically with respect to @xcite . nevertheless we have tentatively detected the h@xmath4d@xmath2 ground - state line ( fig . [ f2 ] ) . the hdo @xmath15=@xmath5@xmath6 ground - state transition ( 464.924520 ghz ) was observed in 1998 july at the jcmt during a night of excellent submillimeter transparency with a zenith optical depth at 225 ghz below 0.05 . the dual channel , dual band 460/660 ghz sis receiver rxw was used in ssb mode with the digital autocorrelator spectrometer ( das ) backend . the das was split in two parts of 125 mhz with a spectral resolution of 189 khz ( = 0.12 km s@xmath10 at 465 ghz ) . the hdo line was observed in the upper side band , and ssb system temperatures including atmospheric losses were @xmath65 k. the beam size is about 11@xmath37 at 465 ghz and the main beam efficiency was measured to be about 50% . we observed the hdo line toward i16293a in position switch mode with the same reference as above . we also observed the line in beam switch mode , switching over @xmath66 in azimuth . these spectra show the same profile as the position switched ones , but only the beam switched spectra clearly reveal the level of the continuum emission from the dust heated by the yso . all spectra were coadded and corrected for the continuum level . in addition , deep spectra of the hdo line toward i16293e were obtained in position switch mode . the data were calibrated similarly to the rxa2 and a3 observations . the h@xmath4o @xmath7@xmath5 ground - state transition ( 556.936002 ghz ) was observed in 1999 august with swas @xcite . the receiver consisted of a schottky diode mixer resulting in a double sideband system temperature of about 2200 k. the back - end was an acousto optical spectrometer with a 1.4 ghz bandwidth and a velocity resolution of @xmath67 km s@xmath10 . the beam width is about @xmath68 . the h@xmath4o line was observed in position switch mode , switching to an emission free reference position 1.5@xmath69 away . the total ( on+off ) integration time was 23.5 hrs , yielding a noiose level of @xmath70 k. the co @xmath15=65 ( 691.473076 ghz ) and 76 ( 806.651806 ghz ) transitions toward i16293a have been taken with the jcmt using the rxw receiver in 1999 august and with the mpifr / sron 800 ghz sis receiver in 2000 april , respectively , under good atmospheric conditions . the beam parameters at these frequencies were determined from observations of mars . the main beam efficiencies are about 0.3 and 0.24 , and the beam sizes ( fwhm ) are @xmath71 and @xmath72 at 691 ghz and 806 ghz , respectively . iras 16293@xmath12422 was found to be a binary system in 6 cm and 2 cm vla observations by @xcite . the two components are separated by 5@xmath37 ( 840 au ) with a position angle of @xmath73@xmath69 . high resolution millimeter observations @xcite also resolve the iras source into a binary system , in good agreement with the vla data . the southern component ( a or mm1 ) is more extended in millimeter continuum and is associated with dense gas , high velocity emission and h@xmath4o masers @xcite suggesting that it is the more active component of the binary system . the northern component , b or mm2 , is very compact both at centimeter and millimeter wavelengths , and has a surprisingly steep spectrum , @xmath74 @xcite . it is brighter than a at both 1.3 cm and 2.7 mm @xcite . we resolve iras 16293@xmath12422 in all maps with a fwhm @xmath75 at 450 with @xmath76@xmath69 , and find it slightly more extended at 800 and 850 with @xmath77@xmath69 and 147@xmath69 , respectively . the corresponding size at 850 @xmath0 m is @xmath78 . the emission is centered on a to within 1@xmath37 in all single dish maps . the same is true for the lower s / n single - coverage ukt14 maps at 1.1 mm , 750 , 450 and 350 . in fig . [ f4 ] , we show the azimuthally averaged flux densities of i16293a and its surrounding envelope , and of the prestellar core i16293e . the intensity falls off much more rapidly for i16293a than for the prestellar core i16293e at radii @xmath79@xmath80@xmath37 , beyond where the slope flattens out to be similar to the prestellar core . the size and @xmath81 seen in our submillimeter maps agree within the errors with the c@xmath82o emission mapped by @xcite with ovro and differs in @xmath81 from the alignment of the two submillimeter sources a and b. in fact , neither the c@xmath82o emission nor the 450 @xmath0 m emission shows any clear enhancement at the b position which is the stronger continuum source in the wavelength range 1.3 cm to 2.7 mm . at longer wavelengths ( 850 and 800 @xmath0 m ) the @xmath81 approaches that of the binary suggesting that at these wavelengths the emission from b is still significant . however , none of our maps lines up exactly with the binary , which is what one would expect if the emission origniated from a circumbinary disk . @xcite claim that their high resolution maps show that the dust emission is aligned with the binary . however , considering the s / n in their map , one could equally well interpret their result as a circumstellar disk surrounding a with a position angle similar to that seen in c@xmath82o and 450 @xmath0 m , and b as a separate but partly overlapping source . their observations show that about 90% of the dust emission surrounding a is spatially extended , while b is largely unresolved at all spatial scales they could measure . @xcite observed the continuum at 1.37 mm with ovro at @xmath83 resolution and find that i16293a and i16293b both have a disk with diameter @xmath84 au . we conclude that the submillimeter dust continuum and most of the molecular emission are centered on a and have a disk - like morphology . the dust disk is surrounded by faint extended emission from the surrounding dark cloud core which falls off rather rapidly toward the south and southwest . the dark cloud core is more extended toward the north and northeast ( @xmath85@xmath37 ) , and in the west a narrow dust bridge connects it to i16293e ( see fig . [ f1 ] ) . in contrast , i16293e is very extended and has a much flatter emission profile than i16293a ( see figs . [ f1 ] , [ f4 ] ) . the peak emission is centered on , with a roughly triangular emission region around the peak . it has a ridge - like structure at 450 @xmath0 m ( see fig . the integrated fluxes ( corrected for error lobe contribution ) are 20 jy and 133 jy at 850 and 450 , respectively , in a 50@xmath37 radius ( 8000 au ) . in order to perform single position photometry we also made gaussian fits to our submillimeter images resulting in a source size of @xmath86 with @xmath87@xmath69 superposed on a more extended background . this source size was used to convert our photometry into integrated intensities for further analysis . the submillimeter maps allow us to derive a mass estimate for the dust which is more accurate than previous values . i16293a has been used extensively as a secondary calibrator for submillimeter continuum observations at jcmt @xcite , and has accurately known fluxes in all submillimeter bands . although some of our ukt14 maps have relatively poor s / n , they still allow us to make a reliable estimate of the flux density of the compact dust disk , because the underlying emission from the envelope can be determined and subtracted . this is not possible from photometry data @xcite . in table [ t1 ] we list the integrated fluxes for i16293a from two - dimensional gaussian fits to our maps after subtraction of the extended envelope emission . we have not mapped the region at 1.3 mm , but instead use data from @xcite who derived the integrated flux in a similar fashion . we could use the scuba data together with co - added iras data at 100 and 60 to constrain the dust temperature , but it is clear from our submillimeter maps that the arcmin size of the iras beam will also include emission from the surrounding dark cloud , so to include the iras data we should do a two component fit and simultaneously solve for the dust envelope and the disk . instead , we take a simpler approach . we solve for each separately and use the iras data to make sure that the results we get are plausible . our deep 850 and 450 scuba maps predict that about half or more of the continuum emission seen by iras is likely to come from the envelope surrounding the dust disk . we therefore divide the iras 100 emission into halves and assume one half to originate in a compact @xmath88@xmath37 disk and the rest from a cooler cloud envelope around the protostar . if we assume that the dust grains can be characterized by a single temperature , @xmath89 , the flux density @xmath90 at frequency @xmath91 can be written as @xmath92 where @xmath93 is the source solid angle , @xmath94 is the planck function , @xmath95 the optical depth of the disk , and @xmath96 optical depth of the envelope . we assume that the envelope is optically thin at submillimeter wavelengths ( @xmath97 ) . we write @xmath95 as @xmath98 where @xmath99 is the dust optical depth at frequency @xmath100 , and @xmath101 is the dust emissivity describing how the dust opacity ( @xmath102 ) changes with frequency @xcite . we adopt the @xcite opacity ( @xmath103 = 0.1 @xmath104 g@xmath10 ) and assume a gas - to - dust ratio of 100 . it is now straight forward to do a least - squares fit to equation ( [ e2 ] ) . since we mapped the whole cloud , we know @xmath93 , and use this to constrain the fit . we omit flux densities measured at 3 mm with aperture synthesis telescopes , because these resolve out the extended emission and may also include free - free emission . for the submillimeter disk of i16293a we derive a dust temperature @xmath105 k ( see fig . [ f6 ] ) , @xmath101=1.6 and a source size of about 5@xmath37 . in order to make sure that these values are not determined by our assumed partitioning of the iras 100 flux estimate , we also repeated the fit omitting the 100 data and got similar results . our fit to the dust disk is not affected by the 100 @xmath0 m data because the dust emission in this source starts to become optically thin in the ( sub ) millimeter regime , @xmath106 0.42 , which acts as a constraint for the dust temperature . this is unusual as dust emission is usually optically thin at ( sub ) millimeter wavelengths and does not provide any constraints on the dust temperature , unless @xmath101 is known . from the fit we derive a total mass of 1.8m@xmath107 corresponding to an average density of @xmath108 @xmath30 , and a bolometric luminosity of 16.5l@xmath107 . the uncertainty in the mass , due to uncertainties in the fitted @xmath89 and @xmath101 , is of the order of 0.3 m@xmath107 . the continuum measurements of @xcite show that the source size is smaller than 3@xmath37 . @xcite assume optically thin emission at 1.37 mm , and a dust temperature of 40 k and derive a lower limit to the mass of 0.25 m@xmath107 . our derived mass of the dust disk is in the mid - range of the values quoted by @xcite . our fit appears to slightly overestimate the flux densities at long wavelengths , which indicates that the data can not completely be described by a single dust temperature . since the continuum emission traces the integrated density along the line of sight , the scuba images can be used to derive the density distribution . we used the azimuthally averaged radial flux density of the continuum maps at 450 and 850 @xmath0 m to determine the radial density structure of i16293e . figure [ f4 ] shows the radial emission profiles of this prestellar core ; it is clearly seen that both emission profiles are well described by a power - law . instead of the dust density we will use the h@xmath4 volume density as parameter in determining the density distribution . we adopt a dust - to - gas ratio of @xmath109 and a power - law h@xmath4 density distribution of the form @xmath110 , where @xmath111 is the h@xmath4 density at the arbitrary chosen radius of 1000 au . similarly , we adopt a dust temperature @xmath89 distribution following a radial power - law with index @xmath112 and temperature @xmath113 at @xmath114 au . the dust emissivity is assumed to be of the form equation ( [ e3 ] ) . the free parameter set of this model ( @xmath115 ) is constrained by the total submillimeter flux , the radial emission profiles , and the spectral index between 450 and 850 @xmath0 m , respectively . a @xmath116 minimalization between the core model emission and the data yields best - fit parameters @xmath117 for both the 450 @xmath0 m and 850 @xmath0 m images , @xmath118 @xmath30 , and @xmath119 . the outer radius of the core was set to 8000 au , the only value of @xmath120 which allows a density profile fit with a single power - law index @xmath121 . the above density structure yields a core mass @xmath122 m@xmath107 . from these results we infer an isothermal dust temperature @xmath123 k , i.e. @xmath124 . we apply this temperature to the region @xmath125 au while for @xmath126 we let the temperature gradually rise to 20 k to include the transition of the dark core to the diffuse ism where the temperature is determined by the interstellar radiation field . this temperature structure gives a good fit to the observed radial emission profiles . note that the continuum emission is a convolution of the density and temperature , so the derived density and temperature structures are degenerate , e.g. , a constant temperature @xmath127 k and @xmath128 @xmath30 also fits the emission profiles . however , the low - end @xmath89 and high - end @xmath27 fit yields a plausible physical description to this prestellar core and best fit to the observed hdo emision ( see below ) . the iras 100 @xmath0 m data were not used to further constrain the above parameter - set , iras was not sensitive to continuum emission from dust with @xmath129 k. on the other hand the scuba emission reflects a convolution of density and temperature along the line - of - sight . thus , a cold high density region with @xmath130 k can easily be hidden in the center of the core since a less dense warmer envelope component will always dominate the observed continuum emission . in general such a region will be hard to detect in continuum emission as well as in line emission . in the latter case one would need a transition of a species with a high critical density and low excitation temperature which is uniquely tracing the coldest and densest part at the heart of the core and is not depleted . deuterated molecules , and in particular the ground - state transitions of hdo , are excellently suited to trace such regions ( see [ s : intro ] and below ) . the inferred radial density and temperature structures from the dust are used to calculate the abundance profiles for the observed single dish observations of c@xmath131o , c@xmath82o , dco@xmath2 , hco@xmath2 , h@xmath3co@xmath2 , and hdo . the modeling of the molecular excitation and line radiative transfer uses a spherically symmetric monte carlo method @xcite . the core was divided in 40 concentric shells . all shells were found to be optically thin at the modeled transitions of the molecules . we assume @xmath132 throughout the core and adopt for each line a constant local turbulent linewidth . after convergence of the level populations was reached , the spectral line profile of the observed transition of each molecule was calculated for the appropriate beam . the calculations were done iteratively starting with an educated guess for the abundance and the turbulent linewidth , until a best fit of the modeled line profiles to the observed spectra was established . our model spectra fit the observed emission profiles for all relevant molecules ( fig . [ f7 ] ) and yield the following abundances for the co - isotopomers : [ c@xmath82o]/[h@xmath4]=@xmath134 and [ c@xmath131o]/[h@xmath4]=@xmath135 . adopting standard co isotopomer ratios @xmath136:[\rm{c}^{18}{\rm o}]=500:1 $ ] , and @xmath137:[{\rm c}^{17}{\rm o}]=2500:1 $ ] @xcite , our derived abundances imply that co is depleted by a factor of about ten with respect to the standard abundance . the width and strength of the observed hdo ground - state emission line is well fit by a [ hdo]/[h@xmath4 ] abundance ratio of @xmath138 . note that the derived turbulent linewidth of the hdo line is much lower than that of the other lines . it is expected that the turbulent line width is roughly the same for all molecular species when observed with similar angular resolution if they are distributed similarly throughout the core . this is in fact the case for the other molecules observed with beams of about 15@xmath37 ( dco@xmath2 , hco@xmath2 , etc . ) for which we obtain an average value of @xmath139 km s@xmath10 . although the weak hdo emission has a s / n of only 3 , its narrowness suggests that it resides in the innermost region of the i16293e core . in this cold region ( @xmath140 k ) the abundance of a deuterated species like hdo is expected to be enhanced through equation ( [ e1 ] ) , while non - deuterated molecules are expected to be depleted ( [ s : intro ] ) . a higher s / n hdo spectrum is required to confirm this . the best fit to the hco@xmath2 32 line yields an abundance [ hco@xmath2]/[h@xmath4 ] of @xmath141 . the h@xmath3co@xmath2 32 emission can be reproduced for a constant abundance of @xmath142 throughout the core , and the dco@xmath2 32 line fit yields an abundance of @xmath143 . note that [ dco@xmath144/[$]h@xmath3co@xmath2]=2.5 , that is the dco@xmath2 abudance is larger than that of h@xmath3co@xmath2 . assuming [ co]:[@xmath3co]=65:1 @xcite , we find that hco@xmath2 is depleted by a factor of about 10 . table [ t3 ] summarizes the inferred abundances and local turbulent linewidths . in this section we determine a radial density and temperature structure for the yso envelope using an inside - out collapse power - law density structure where the slope of the density and the velocity depend on the location of the collapse expansion wave . we assume that the dust temperature follows a radial power law . such a distribution is expected for a spherical cloud with an embedded heating source in its center . together with the fir luminosity and distance of 160 pc , the temperature distribution is constrained from the total observed submillimeter continuum emission from the disk and the envelope ( table [ t1 ] ) . the inferred temperature profile ranges from @xmath145 k at the inner radius @xmath146 au , to @xmath147 k at the outer radius @xmath148 au . we use the spherically symmetric self - similar solution of a collapsing cloud core derived by @xcite to determine the density profile . this model is characterised by only two parameters : the sound speed @xmath9 , and the time @xmath11 since the collapse starts at the center at @xmath149 . the initial stationary density distribution is an isothermal sphere where the density is proportional to @xmath150 . at a time @xmath11 after the onset of collapse , the head of the collapse expansion wave is radius @xmath151 . inside this radius the velocity field varies from stationary to free fall , @xmath152 , while the density varies as @xmath153 . we determine the infall parameters independently from the molecular spectroscopic data as well as from the dust continuum measurements . we start with the former . we do not use rxa2 observations , which instrumentally broadens the lines , but only 345 ghz window spectra to determine the infall parameters . earlier studies often used emission lines from lower rotational transitions of hco@xmath2 or h@xmath3co@xmath2 . our line profiles of hco@xmath2 43 and h@xmath3co@xmath2 43 are complex with excess emission at the red and blue wings of the doubly peaked emission profile , respectively ( see below ) . this points to a contribution by the outflow(s ) . the red and blue wings of the dco@xmath2 54 spectrum are symmetric , and we therefore use this line profile to constrain the parameter set @xmath9 and @xmath11 . these parameters respectively determine the width and integrated line intensity of the profile . a good fit to the wings of the dco@xmath2 54 emission is found for @xmath154 yr , @xmath155 km s@xmath10 , and a constant radial abundance [ dco@xmath2]/[h@xmath4]=@xmath156 ( fig . this yields an envelope h@xmath4 mass @xmath157 @xmath158 within a radius of 7300 au . note that the absorption can easily be fitted by adding a cold outer shell where the temperature has dropped to @xmath159 k. our derived stationary envelope mass corresponds well with the mass derived by @xcite of @xmath160 inside @xmath161 au , but is as much as a factor of four higher than the masses derived by @xcite ( @xmath162 , @xmath163 au ) and @xcite ( @xmath164 , @xmath165 au ) . in fig . [ f9 ] we plot the calculated dust and density profiles . for comparison we also plot the independently derived density and temperature profiles from @xcite and @xcite on a log - log scale , so power - law distributions become straight lines and the outer regions of the envelope stand out more clearly . our profiles agree well with those of @xcite who derived the density and temperature profiles independently from a subset of our continuum measurements . our values for the infall parameters compare well with those derived by @xcite ( @xmath166 km s@xmath10 , @xmath167@xmath168 yr ) and @xcite @xmath169 yr ) . the age is more than a factor of two lower than derived by @xcite . this is probably because our study is focused on high-@xmath15 transitions , and therefore more sensitive to the warm inner envelope regions , while previous work dealt mostly with lower @xmath15 transitions which are dominated by the gas in the colder outer envelope regions . we used the azimuthally averaged radial emission profile ( fig . [ f4 ] ) to determine independently the infall parameters from the scuba continuum maps . a best fit of the inside - out collapse model from @xcite yields a sound speed @xmath155 km s@xmath10 and an age @xmath170@xmath171 yr , close to the values derived from our fits to the line profiles . this indicates that the model of @xcite can be used to predict the infall velocities for this class of ysos , but that the age is better constrained from the line profiles . @xcite also found a discrepancy between the infall parameters derived from dust continuum and molecular line emission . the infall analysis of @xcite for a sample of yso envelopes indicate that this discrepancy may be a common property of class 0 ysos , since a better correspondence was found for class i and older sources . a possible explanation is that collapse models generally predict a much higher mass - accretion rate early in the evolutionary stage than the @xcite model does . we therefore use the infall parameters @xmath9 and @xmath11 from the dco@xmath2 fit to model the profiles of the observed molecular lines and to infer the abundances , the results are presented below . the best fit to the observed hdo profile implies a constant abundance of @xmath172 throughout the envelope . the model profile is narrower than observed this may indicate a significant contribution to hdo emission from , e.g. , the outflow(s ) ( [ outflow_model ] ) . to reproduce the line wings by infall , the sound speed must increase to @xmath173 km s@xmath10 . the fit to the absorption could be improved by adding a cold shell with @xmath174 k. but , hdo collision rates are only available for @xmath175 k @xcite . our models use the 50 k rates at lower temperatures while the true rates may easily be a factor of two lower at @xmath1761215 k enhancing the optical depth of the hdo line . we therefore expect that extended collision rates would improve the fit of the line profile . the continuum level results from the included fir radiation field in the excitation analysis and matches the observed level very well ( fig . [ f8 ] ) . we use the best hdo infall model fits ( @xmath173 km s@xmath10 , @xmath177 yr ) to model the hdo @xmath60@xmath59 and @xmath58@xmath59 lines . these lines have energy levels far above ground and are not expected to be collisionally excited for this temperature range . however , in high mass ysos these lines are found to be much stronger than expected on basis of collisional excitation @xcite , probably because the high levels are populated through excitation by the fir radiation field . with the earlier mentioned dust parameters for the fir excitation , the model yields a @xmath60@xmath59 peak intensity of @xmath178 k , and @xmath58@xmath59 @xmath179 k. the former value is much lower than the upper limit to the line strength of 0.08 k from @xcite for the @xmath60@xmath59 line . since the upper - limits of @xcite for the @xmath60@xmath59 and @xmath58@xmath59 lines are high with respect to what can be achieved with present - day , more sensitive receivers , we decided to reobserve both lines . deep spectra reveal no features at a noise rms of 0.02 k at either frequency for a spectral resolution of 1.25 mhz ( = 1.7 km s@xmath10 ) . the excitation of h@xmath4o has been calculated using the collision rates of ortho- and para - h@xmath4 with ortho - h@xmath4o @xcite for kinetic temperatures between 20 and 140 k for the lowest 5 rotational energy levels of h@xmath4o . this range traces the temperature structure of our model very well . we interpolated the rates for the appropriate temperature of each shell . the ortho / para h@xmath4 ratio in each shell is assumed to be in lte at the appropriate temperature of each shell ; for low temperatures , @xmath180 k , most of the h@xmath4 is in its para modification and [ ortho - h@xmath4]/[para - h@xmath4]@xmath181 for lower @xmath182 . we included fir dust excitation in our model , using the dust emissivities from @xcite . calculations with a large range in power - law emissivities show that the dust emission has not much impact on the excitation of the lowest water lines in the envelope of i16293a . therefore , we did not consider the fir radiation in our further calculations . the total h@xmath4o abundance is derived assuming an ortho - to - para h@xmath4o ratio of @xmath183 . the abundance was varied to fit the self - reversal and to match the @xmath7@xmath5 peak h@xmath4o emission observed by swas . our infall models fail to match the large width of the observed h@xmath4o spectrum . this is not surprising ; in the large swas beam ( fwhm@xmath184 arcmin @xmath185 au ) emission from our model cloud is dominated by the cold outer envelope . the width of the observed line profile clearly indicates that most of the h@xmath4o emission arises in molecular outflows associated with this class 0 yso which are not included in our envelope model . we therefore focus here on the region causing the absorption and the emission at the systemic velocity of the cloud , and model the h@xmath4o outflow in [ outflow_model ] . our envelope model predicts a narrow double - peaked profile which can match the intensity level of the observed maxima , but not the width ( fig . [ f10 ] ) . initially , the analysis was done for a uniform abundance [ ortho - h@xmath4o]/[h@xmath4]=@xmath186 across the envelope . this fits the maximum emission , but causes too deep an absorption . to reduce the absorption we applied a step function for the abundance profile with a drop in the h@xmath4o abundance at @xmath187 k , where atomic oxygen starts to freeze out . such a two - step abundance profile fits the observations for [ ortho - h@xmath4o]/[h@xmath4]=@xmath188 for @xmath189 k ( fig . [ f10 ] ) . the fir h@xmath4o lines observed with iso @xcite are much stronger than predicted by our model , indicating that they do not originate in the envelope . the h@xmath4d@xmath2 @xmath7@xmath8 line has been marginally detected at a 3-sigma level . a constant abundance of @xmath190 yields a good fit to the observed maximum emission , but the line width of the model spectrum is too large . we therefore tried a step function for the abundance , [ h@xmath4d@xmath2]/[h@xmath4]=@xmath191 for @xmath192 20 k and [ h@xmath4d@xmath2]/[h@xmath4]=@xmath193 for @xmath194 k which yields a good fit to the complete line profile ( fig . [ f8 ] ) . the n@xmath4h@xmath2 43 line was observed simultaneously with the h@xmath4d@xmath2 @xmath7@xmath8 line . a constant abundance with the radius of @xmath195 yields a good fit to the observed line profile ( fig . [ f8 ] ) . we can only put a lower limit to the hco@xmath2 abundance of @xmath196 , where the line 43 starts to become optically thick . however , this does not match the peak intensities well , nor does it fit the apparent extra blue and red components . the wings are well matched for @xmath173 km s@xmath10 . an outflow is likely contributing to the emission . the abundance for @xmath197 10 k could be enhanced to get a deeper absorption , but this would not add much information since the emission and the absorption are already optically thick . note that the high optical depth of the emission and absorption masks any signature of chemical changes at low temperatures . this transition is therefore not suited for an accurate derivation of the abundance structure . for this hco@xmath2-isotopomer , we also have problems in modeling the observed double line structure , probably because kinematically distinct components are present at the blue side ( fig . [ f8 ] ) . an abundance of @xmath198 gives a good fit to the red component . this component is optically thin and we attribute it to the envelope . note that a constant abundance thoughout the envelope is used for this fit . the blue component is probably associated with one of the quadrupole outflow components ( [ co_outflows ] ) . the co 65 and 76 lines have been modeled for a @xmath61co abundance [ @xmath61co]/[h@xmath4]=@xmath199 . in fig . [ f11 ] the model spectra are overlaid on the observations . it can be seen that the envelope model fails to fit the observed emission for a sound speed @xmath155 km s@xmath10 . even for @xmath173 km s@xmath10 , the width and maxima of the observed spectra can not be matched . this indicates that , like the h@xmath4o emission , most of the observed high-@xmath15 @xmath61co emission is associated with the outflows ( [ outflow_model ] ) . the c@xmath82o 32 line profile can well be fitted for an abundance of @xmath18 throughout the envelope , corresponding to a ` standard ' [ co]/[h@xmath4 ] abundance of about @xmath200 . @xcite derive a c@xmath82o abundance of @xmath201 on basis of similar observations with the jcmt . however their abundance yields a fit to the spectrum which is about a factor of two lower than our fit ( see their fig . 4 ) , so the agreement is within a factor of 1.5 . our abundance - fit yields a c@xmath82o 21 spectrum whose central region is quite different from the observations indicating a lower abundance . the observed low emission level of the c@xmath82o 21 may be caused by a low - density c@xmath82o component with a ` standard ' abundance residing in dark / translucent transition region between the outer envelope and the diffuse interstellar medium , while co in the cold outer envelope regions may be highly depleted . moreover , the modeled c@xmath82o 21 and 32 line profiles are somewhat narrower than observed . the line wings of both transitions can be fitted by infall if a sound speed of @xmath202 km s@xmath10 is adopted , or there may be a contribution from the outflow . figure [ f8 ] plots the excitation model fit for the c@xmath82o 32 line . we conclude by noting that the low-@xmath15 c@xmath82o transitions are not well suited to determine the co depletion in the envelope since the line profile indicates than c@xmath82o is sensitive to material which may not reside in the envelope . we use a constant c@xmath131o abundance of @xmath203 , i.e. , a ` standard ' co abundance for @xmath204 k and depletion by a factor of 50 for @xmath205 k. this fits better than a uniform abundance throughout the envelope , because depletion makes the line appear broader . the line is also broadened by unresolved hyperfine lines , e.g. , visible as a low - velocity shoulder . still , the observed line is broader than our model predicts . the wings fit better for a sound speed @xmath202 km s@xmath10 . there is a hint for a depression of the peak emission . @xcite derive a c@xmath131o abundance of about @xmath206 which agrees within a factor of two with our results using a constant abundance value . the ortho - h@xmath38o@xmath2 @xmath207@xmath208 line at 396.272412 ghz has been searched for in i16293a by @xcite . we used their upper limit and our envelope excitation model to constrain the abundance [ ortho - h@xmath38o@xmath144/[$]h@xmath4]@xmath209 . this correspondes to an [ h@xmath38o@xmath144/[$]h@xmath4 ] upper limit of ( 35)@xmath210 if the h@xmath38o@xmath2 resides mainly in the @xmath211 k region , or @xmath212 k region , where the [ ortho - h@xmath38o@xmath2]:[para - h@xmath38o@xmath2 ] ratio varies between @xmath213 and @xmath214 , respectively . these upper limits are about an order of magnitude larger than derived by @xcite using statistical equilibrium calculations for a slab with a homogeneous density and temperature . the derived abundances of all species are summarized in table [ t3 ] together with those of @xcite . taking into account the differences between their fits to the observed spectra and ours ( see above ) , the abundances agree within a factor of 1.5 . the symmetry axis of the ne sw outflow , as determined from our co 21 map , goes through a at all velocities ( fig . [ f12 ] ) . the centroid of the overlapping blue- and red - shifted emission is centered on i16293a , which is also evident from the velocity position plot in fig . [ f13 ] ( although inescapably includes some of the e w outflow ) . furthermore , aperture synthesis maps of dense outflow tracers , especially cs @xcite and so @xcite clearly show high velocity gas emanating from i16293a in the direction of the co outflow , while i16293b shows no activity , nor does it seem to be associated with a maximum in any high density gas . we therefore attribute the ne sw outflow to i16293a . our submillimeter maps ( figs . [ f1 ] and [ f4 ] ) clearly show that the dust emission associated with the free - free and millimeter - source i16293a is extended . this is one of the few ysos where we can resolve the dust emission : it has a disk - like structure with a size of roughly 2000 au . at low frequencies the continuum emission from i16293a splits up into a double source . we do not believe that i16293a is in itself a binary ; a more plausible interpretion is that we are seeing highly collimated ionized gas from the outflow , as has been seen in l1551 irs5 @xcite , where the jet dominates the emission at low frequencies and the disk emission becomes dominant at high frequencies . these are strong arguments for an accretion disk that is centered on i16293a and drives a powerful outflow , which is very well collimated close to the star ( e.g. , free - free emission ) , but relatively highly collimated even in co. the ratio of length to width is about 1:3 at high velocities . the outflow is at high inclination , since we can see overlap between the blue- and the red - shifted lobe . if we assume the outflow to be conical , which according to the velocity position plot ( fig . [ f13 ] ) appears to be a reasonable assumption , we can use the method outlined by @xcite to estimate the inclination of the outflow , and find about 65@xmath69 for both the blue and the red lobe . this is slighly higher than what @xcite find on the basis of their sio 21 map for the ne red lobe , where they derive an inclination by 4045@xmath69 from the plane of the sky . the properties of the outflow have been derived in several studies ( e.g. , @xcite , but we will redo it here ; partly because we have more accurate data , but also to point out how uncertain these estimates are . the deep integrations of @xmath3co and c@xmath82o toward i16293a ( fig . [ f2 ] ) clearly show high velocity emission , which indicates that the @xmath61co optical depth in the wings is substantial . at near outflow velocities ( @xmath215 1.5 km s@xmath10 from the cloud systemic velocity ) we find optical depths of 40 or more , both from @xmath3co and c@xmath82o , but note that here the line wings originate from both outflows and also get a significant contribution from the underlying accretion disk . we therefore obtained additional @xmath61co and @xmath3co spectra at the peak positions in the lobes of the ne sw outflow , i.e. , at offset positions ( + 80,+55 ) for the redshifted lobe and ( @xmath1130,@xmath140 ) for the blueshifted lobe . these show that the red - shifted outflow has substantially higher optical depth . we estimate a @xmath61co 21 optical depth of about unity at radial velocities as high as 11 km s@xmath10 away from the systemic velocity of the cloud , whereas the optical depth has dropped to unity at + 6 km s@xmath10 away from the cloud core for gas in the blue - shifted outflow . the @xmath3co spectra are not sensitive enough to enable us to derive optical depths of less than about unity . we compute the outflow mass , momentum , momentum flux , and energy content with and without opacity corrections , by assuming that we can apply the same opacity correction to all of the gas in the outflow . since we have not directly measured the excitation temperature of the outflowing gas we assume it to be 50 k , but note that it could be as high as 80 k @xcite . the values in table [ t4 ] are obtained assuming an abundance ratio 10@xmath22 for [ co]/[h@xmath4 ] and summing over the map in 1 km s@xmath10 wide velocity intervals starting 1 km s@xmath10 from the cloud systemic velocity ( assumed to be 4 km s@xmath10 ) . the apparent dynamic outflow time scales are about the same for both lobes , ( 3.03.5)@xmath216 yr . this agrees with the dynamical timescale of ( 57)@xmath216 yr derived by @xcite for the ne red lobe . however , the blue outflow lobe appears more energetic . normally when one sees an imbalance in the momentum or force of the outflow , it is an indication that one of the outflow lobes has penetrated the cloud surface and therefore has no gas to interact with . here we see no evidence for this . a more likely explanation is that we have not properly corrected for the opacity and excitation differences in the two outflow lobes . the blue lobe is more extended and has higher apparent velocities , suggesting that the gas surrounding the blue lobe is less dense than the gas surrounding the red outflow lobe . this is also seen in our opacity estimates , which are higher for the red - shifted outflow . a similar conclusion is reached by @xcite and @xcite on basis of sio and ch@xmath38oh maps , respectively . the symmetry axis of the e w outflow passes north of i16293a , but not as clearly defined as the ne sw outflow . the red - shifted outflow lobe is much more compact than the blue - shifted lobe , and appears to bend more towards the north . however , the peak intensity of the outflowing red - shifted gas is rather well aligned with the large blue - shifted outflow lobe with an average position angle for the flow of @xmath217@xmath69 . the e w outflow appears to intersect with the continuum disk about 510 north of i16293a . since there is no source near this position other than i16293b , we associate the e w outflow with i16293b . a similar conclusion was also drawn by @xcite , although they stress that at present there is no evidence for outflow activity in the immediate vicinity of b. @xcite find that the ch@xmath38oh wing emission has a narrow and nearly constant velocity width along the e w lobes , which they attribute to turbulent entrainment . this is consistent with our finding that the e w outflow is a fossilized flow . it has been suggested that the ne red outflow lobe and the eastern blue outflow lobe could be part of a wide angle flow similar to the l 723 outflow @xcite or that the two outflows are part of a precessing or episodic outflow driven by the same star @xcite . however , our co map clearly shows that these are two separate outflows . since the symmetry axis of the e w outflow does not intersect with i16293a , it is rather unlikely that this outflow would have been caused by an earlier outflow episode from i16293a . @xcite and @xcite also conclude that two independent bipolar outflows are present on the basis of their sio and ch@xmath38oh maps . @xcite also discuss a third outflow , because in their low spatial resolution nagoya map , the e w outflow extends more than 10 from iras 16293@xmath12422 and their high resolution nobeyama map does not show this extended outflow component . our map does not extend as far as the nagoya map , and our spatial resolution is poorer than that of the nobeyama co 10 map . however , the co map we present here has the advantage of being fully sampled , unlike the sio 21 emission mapped by @xcite , and it has a much higher resolution ( 20@xmath37 ) than the large - scale sio 21 map of @xcite ( @xmath218 ) . we therefore better resolve the two outflows . one can clearly see that the outflow continues to the east outside the area we have mapped , but with rather low velocities . therefore the extended blue low - velocity outflow appears simply a continuation of the e w flow . we also clearly see the interaction of this outflow with the i16293e core , which was already found by @xcite . in our channel maps , which are overlayed on the 800-continuum emission ( fig . [ f12 ] ) , this interaction is clearly visible . it appears rather gentle : the outflow is enhanced at low velocities and seems to stream around the core and forms almost a cavity immediately behind it . that it streams around the core region is supported by the fact that we see faint red - shifted @xmath61co towards the dense core region , but the red - shifted emission is much stronger in front of the core , although one can still see it faintly on the backside as well ( fig . [ f12 ] ) . figure [ f5 ] shows a blow - up of the 450 @xmath0 m image of i16293e . it is clear that the outflow has compressed the core and caused the high density n s ridge that we see in our continuum maps and prominently at near cloud velocities in the channel - maps of fig . [ f12 ] . behind the core it appears that the outflow gets recollimated . there is no indication of interaction between i16293e and the red - shifted ne outflow lobe from our co channel maps , nor from the sio channel maps @xcite . @xcite attribute the blue and red stream of the e w outflow around the i16293e core to the red and blue lobe of an outflow powered by i16293e causing two shocked regions ( he1 and e1 in their nomenclature ) visible in their h@xmath4co maps . they support their hypothesis by means of hco@xmath2 10 spectra taken along a slice which goes through i16293e as well as through the peaks in the h@xmath4co emission , since these spectra reveal a blue wing at one side of i16293e and a red wing on the counter side . although these wings are real , the results are different when put in context of the entire cloud . our large maps show what small maps , which do not fully cover the complete interaction region of the extended outflows with the ldn 1689n cloud , can not : local small scale flows in the vicinity of i16293e are part of large scale motions . from our co map it is clear that the apparent sw ne outflow and the alignment with i16293e are due to projection effects of pieces of the e w outflow from i16293b which streams around the dense prestellar core i16293e . this view is supported by the study of @xcite who find that the e w flow has fan - shaped lobes , each of which containing blue- and red - shifted components . this would imply that e1 is red - shifted gas in the far side of the eastern blue lobe . the red - shifted shocked region e1 is thus the interaction region of the e w outflow , while the blue - shifted h@xmath4co peaks he1 and he2 from @xcite are the interaction of the near side of the e w blue lobe with dense gas of ldn 1689n located at the southwest and southeast side of the dense core i16293e . these interaction regions could also be the edges of i16293e . section [ envelope ] found that the envelope model of i16293a can reproduce the line profiles of most observed molecular transitions , but not of co 65 and 76 , hdo , and h@xmath4o . although the absorption at the cloud systemic velocity can be fit with an envelope model , the broad emission can not be reproduced even for sound speeds as high as @xmath173 km s@xmath10 . here we investigate the origin of the broad high-@xmath15 co , hdo , and h@xmath4o wings by means of an outflow model . we have scaled our co 76 emission to the level of the h@xmath4o emission ( fig . [ f14 ] ) . it is striking that the two spectra have such a similar shape despite the fact that the beam sizes differ by a factor of about 30 . the match of the absorption widths is also very good , and indicates that we may see here the same region with jcmt and swas in absorption . in addition , the line wings are similar , although at the blue side there is an extra co emission component while at the red side there is an extra h@xmath4o emission component present . but the extent of the wings is similar for both species . does this mean that the wings are from a very small region covered by both beams , or that the wing component is rather uniformly distributed throughout the cloud ? the swas data were fitted with a monte carlo radiative transfer code and a physical model of a spherically symmetric cloud subdivided into shells of constant physical conditions . the use of the code and its application to water lines is described in @xcite , which also analyzes the variety of water line profiles and establishes broad diagnostics of outflow , infall , and turbulence apparent in swas data . the line shape lead us to explore models which combine outflow and turbulence ; we find that the line strengths depend on the velocity field as well as on the temperature and abundance in a complex relation . figure [ f10 ] ( middle panel ) shows the predicted ortho - h@xmath4o spectra from the outflow model with an ortho - h@xmath4o abundance of @xmath219 throughout , a turbulent velocity of 2.3 km s@xmath10 , and maximum outflow velocity @xmath220 km s@xmath10 . the temperature ranges from 200 k to 15.5 k. the outflow model provides an acceptable fit to the observed extended wing emission , but can not explain the central part of the observed profile . the outflow model results are complementary to the infall modeling of the ortho - h@xmath4o line ( fig . [ f10 ] , top and middle panel ) , which fits only the central region of the observed spectrum . to illustrate this qualitatively , we plotted in fig . [ f10 ] ( bottom pannel ) the sum of the outflow and envelope spectra of ortho - h@xmath4o which indeed yields a rather good description of the entire swas spectrum . note that the red wing of the h@xmath4o emission is not fitted by our model , this may be due to multiple outflows ( [ nesw ] and [ ew ] ) which we did not model . for a quantitative comparison we do not only need a mutiple ouflow model but also have to incorporate the outflow and infall models in one single excitation model since they overlap in extent and velocity space . this could mean that part of the blue infall gets absorbed by the blue outflow component and vice versa . an avenue for further study would be a combined infall and outflow model . the same outflow models were used to model the co 65 and 76 lines in the jcmt beam using an abundance [ co]/[h@xmath4]=@xmath200 ( fig . [ f11 ] ) . the wing emission predicted by the outflow model poorly fits the observed emission which indicates the presence of multiple outflows , analogous to the h@xmath4o situation . finally , we calculated the hdo emission in the outflow in a @xmath221 beam . we started with a uniform [ hdo]/[h@xmath4 ] abundance of @xmath222 , similar to the envelope value . this yields a spectrum which matches the maximum intensity of the observed emission , but not the wings ( fig . [ f15 ] ) . this may be due to the presence of multiple outflows , similar to the h@xmath4o case . the outer wings can be fitted for an abundance of @xmath16 and results in a self - reversed profile , but its peak intensity is too high by a factor of 24 . this is not surprising since we do not include an infalling envelope . a better fit may be obtained by a combined ( multiple ) outflow and infall model as discussed above for h@xmath4o . we conclude that our shock model can explain the wings of the observed hdo emission and that the hdo abundance in the outflow may be somewhere in the range from @xmath190 to @xmath223 . it is clear that great caution should be taken when comparing the h@xmath4o abundance derived from the swas measurements and the hdo abundance derived from the jcmt observations . nevertheless we are able to use the same models since we convolve the emerging emission with the appropriate beam . because of the large beamsize , swas is not sensitive to the h@xmath4o emission from the warm inner envelope . to test this we enhanced the abundance of h@xmath4o for @xmath224 90 k to the value of @xmath200 this was not found to have any significant impact on the emission in the swas beam . clearly swas is mostly sensitive to h@xmath4o emission from extended colder gas in the outflow , and to a lesser extent to gas in the outer envelope region ( [ outflow_model ] ) . the shocks associated with the outflows are able to release h@xmath4o from the grains and effectively convert oxygen into h@xmath4o through neutral - neutral reactions , but the velocities are so small that h@xmath4o formation at an @xmath225 abundance level is inhibited . indeed , iso lws observations yield much lower h@xmath4o abundances . the iso observations of h@xmath4o are all from higher level transitions @xmath226 k , and thus trace the hot / warm disk / core region of small extent . @xcite derive an [ h@xmath4o]/[h@xmath4 ] abundance in the ` outer envelope ' ( @xmath227 au ) of about @xmath228 which is similar to our envelope abundance ( @xmath229 ) , while for the inner region ( @xmath230 au ) they derive an enhanced abundance of @xmath231 which is attributed to evaporation of water ice mantles of grains at @xmath232 k. the latter value is higher than our inferred outflow abundance of about @xmath20 ( [ outflow_model ] ) . if we combine the h@xmath4o abundances of @xcite with our derived hdo abundances we find that the [ hdo]/[h@xmath4o ] abundance ratio varies between @xmath233 in the warm envelope , @xmath234 au , @xmath235 in the envelope with @xmath236 au and @xmath237 k , and @xmath238 in the outer envelope where @xmath239 k. for a detailed comparison between the hdo and h@xmath4o abundances throughout the envelope we consider in the next section the temperature effects of ion - molecule chemistry and the effects of depletion of molecules and atoms which play a vital role in the formation of hdo and h@xmath4o . in ion - molecule chemistry ( e.g. , @xcite ) hdo is thought to be formed through the dissociative recombination @xmath240 with a statistical branching ratio of compared to the corresponding recombination of h@xmath38o@xmath2 . in steady state the gas - phase density of h@xmath4do@xmath2 can be written as @xmath241 where @xmath242 are the relevant reaction rate coefficients . we adopt equal total rate coefficients for the reactions involving the deuterated and hydrogenated species and assume statistical branching ratios . h@xmath38o@xmath2 is formed through similar reactions with hydrogenated species . we will neglect the formation of h@xmath38o@xmath2 from channels leaving the h@xmath4do@xmath2 reaction sequence and vice versa , and neglect the formation of h@xmath38o@xmath2 and h@xmath4do@xmath2 from reactions of h@xmath46 or h@xmath4d@xmath2 with hco@xmath2 , and dco@xmath2 , respectively . the [ hdo]/[h@xmath4o ] abundance can then be written @xmath243 the heart of the fractionation of deuterated molecules is the molecular - ion h@xmath4d@xmath2 which is formed through reaction ( [ e1 ] ) . in equilibrium the formation and destruction rates of h@xmath4d@xmath2 are balanced which leads to @xcite @xmath244 where @xmath245 is the dissociative recombination rate coefficient of h@xmath4d@xmath2 , @xmath246 is the rate coefficient for the formation of h@xmath4d@xmath2 via the reaction h@xmath247 , @xmath248 is the fractional abundance of species x = co , o , h@xmath4o , etc . , and @xmath249 and @xmath250 are the forward and backward rate coefficients of reaction ( [ e1 ] ) , see @xcite for references of the adopted rate coefficients . the h@xmath46 density depends on the effective cosmic ray ionization rate @xmath251 and can be written as @xcite @xmath252 we consider only the destruction of h@xmath4d@xmath2 through reactions with co , assume @xmath253 , and @xmath254 s@xmath10 . together with the balance for h@xmath46 formation and destruction ( eq.[[e9 ] ] ) we derive an expression for the h@xmath4d@xmath2 abundance with the co abundance as the only free parameter . this abundance structure was subsequently used as input in the excitation calculations to compute the spectrum of the ortho - h@xmath4d@xmath2 ground - state transition . a good fit to the observed h@xmath4d@xmath2 spectrum is established for a two step co abundance profile with [ co]/[h@xmath4]=@xmath255 for @xmath256 22 k and @xmath257 for @xmath258 k ( fig . [ f16 ] ) applied to the infall envelope model . this indicates a co - depletion by a factor of 50 at temperatures below 22 k , where co is frozen out . this value equals the depletion derived from our c@xmath131o 21 excitation analysis , but is a factor of 50 larger than derived from our c@xmath82o 21 and 32 observations which indicated no depletion ( [ envelope ] ) . however , the spectra of the latter transitions are biased toward the warm inner regions and the outflow . now we can use the [ h@xmath4d@xmath144/[$]h@xmath46]-model fit , together with the h@xmath4o abundances from the excitation analysis and the expression of the [ hdo]/[h@xmath4o ] ratio ( eq . [ [ e7 ] ] ) to determine an hdo model abundance profile and to calculate a model spectrum analogous to h@xmath4d@xmath2 . the model spectra are shown in fig . when we adjust [ h@xmath4o]/[h@xmath4]=@xmath260 for @xmath261 k and @xmath193 for @xmath189 k we get a fit to the observed emission peaks . note that these abundances agree within the uncertainties with the derived h@xmath4o excitation results . the hdo chemistry model spectrum ( fig . [ f16 ] ) resembles the hdo excitation infall spectrum ( fig . [ f8 ] ) very well . the fact that the absorption does not go down to zero is a result of the uncertainty in the hdo collision rates for @xmath262 k. note that the hdo spectrum is calculated self - consistently from the theoretical [ hdo]/[h@xmath4o ] expression and the h@xmath4o 2-value abundance profile from the excitation analysis , i.e. , it does not provide an independently determined hdo abundance structure . the upper - limits on the ortho - h@xmath38o@xmath2 abundances derived from the observations by @xcite in [ envelope ] can be used to determine an upper limit to the h@xmath4o abundance . from the formation and destruction balance of h@xmath4o , we derive [ h@xmath4o]/[h@xmath38o@xmath2]@xmath2631500 where we used the experimentally determined branching ratio of 0.5 from @xcite to form h@xmath4o in the dissociative recombination of h@xmath38o@xmath2 . furthermore we only considered the formation of h@xmath4o via the h@xmath46 channel and assume that 10% of all ions cause destructive dissociation of h@xmath38o@xmath2 , similar to @xcite . the upper- and lower - limits correspond to the @xmath264 k and @xmath265 k ranges , respectively . the observations then yield [ h@xmath4o]/[h@xmath4]@xmath266 for both temperature ranges and are consistent with the abundances derived by @xcite from iso observations . the derived h@xmath4o abundance in the outflow is @xmath267 and agrees with the dynamical / chemical model of @xcite for a pre - shock chemistry which evolves until @xmath268 yr . the same model predicts significant deuterium fractionation in the pre - shock gas after @xmath269 yr : [ hdo]/[h@xmath4o]@xmath270 , in agreement with our derived value for the envelope region where @xmath236 au and @xmath271 k. for the warm inner - envelope region @xmath272 au , we derive [ hdo]/[h@xmath4o]=@xmath233 . our gas - phase chemistry model predicts values about two orders of magnitude lower for this region . we attribute the observed enhancement to the shock induced release of solid water ( hdo , h@xmath4o ) from the grains . our derived [ hdo]/[h@xmath4o ] ratio in the cold ( @xmath17 k ) envelope regions is @xmath273 . this ratio is however an upper limit due to the fact that the hdo collision rates are only available for @xmath274 k. for lower temperatures we used the 50 k values , but the rates are expected to decrease for lower temperatures , so that the true hdo abundance may be lower than we derived in [ co_outflows ] . since oxygen starts to freeze out at @xmath187 k we may also expect a drop in hdo which is similar to h@xmath4o ( [ simple_chem_model ] . thus , the [ hdo]/[h@xmath4o ] ratio may be in the range @xmath235@xmath238 at @xmath189 k. the [ hdo]/[h@xmath4o ] chemistry model ( eq . [ [ e7 ] ] ) reaches a maximum value of @xmath275 when [ h@xmath4d@xmath2]/[h@xmath46]@xmath276 . thus , the observed [ hdo]/[h@xmath4o ] ratios can be understood from pure gas - phase deuterium chemistry at low temperatures . the [ hdo]/[h@xmath4o ] abundance ratio in the warm inner - envelope @xmath277 of i16293a is close to the value that has been measured in the solar system : @xmath278 in the comets hale - bopp , halley , & hyakutake @xcite , and @xmath279 in the earth s oceans @xcite . these values are all so close as to suggest that the [ hdo]/[h@xmath4o ] abundance ratio may be a constant in the cycle of star- and planet formation . the fundamental question is then : where and when is the [ hdo]/[h@xmath4o ] ratio determined ? there are currently two favoured scenarios ( e.g. , @xcite ) : 1 ) deuteration from ion - molecule chemistry in the protoplanetary disk ; 2 ) deuteration determined by gas phase and/or solid state chemistry in the dense interstellar medium , namely in the parent molecular cloud before star formation begins . @xcite modeled the [ hdo]/[h@xmath4o ] ratio in accretion disks models and found values of the order of @xmath235 for @xmath280 au in the cold midplane where @xmath281 k. this may indicate a large depletion of h@xmath4o while hdo is still in the gas phase . unfortunately the jcmt and swas obervations can not be used to determine the ( deuterated ) water abundances in the circumstellar disk . for the moment we can only make a comparison for the hdo abundance in the warm envelope . our derived [ hdo]/[h@xmath4 ] abundance ratio of @xmath190 is comparable within a factor of a few to the protoplanetary disk model of @xcite with mass accretion rate @xmath282 yr@xmath10 , and supports the hypothesis that the hdo abundance in the disk and warm envelope are equal . i16293a is a class 0 source , i.e. , is believed to be a protostar according to the classification proposed by @xcite . it is invisible in the near- and the mid - infrared , it is associated with free - free emission so it has already formed a stellar - like core , it drives a powerful molecular outflow , and it is surrounded by an envelope of dust and gas . even though the temperature that we derive for the dust disk surrounding i16293a is higher than the canonical value of 30 k proposed by @xcite , the ratio of submillimeter to bolometric luminosity is typical for class 0 sources , namely @xmath283 . for i16293a we derive @xmath284 . note that this is really a lower limit , since our estimate for the bolometric luminosity also includes the luminosity of i16293b , which is likely to be a few l@xmath107 , while its contribution to the submillimeter luminosity is negligible . @xcite , in their study of outflow properties of a sample of class i and class 0 sources , find that class 0 sources in general have a much higher outflow efficiency ( defined as the ratio of momentum flux to radiative force ) than class i sources , with an average ratio for class 0 sources of about 1000 . for the i16293a outflow we find @xmath285 , which is close to the average value found by @xcite . they also included i16293a in their sample , but the outflow momentum they derive is significantly larger because they also include the e w outflow , which is driven by i16293b . although it is most likely that i16293a and i16293b originate from the same collapsing core , i16293b remains much more of an enigma . our observations support the suggestion of @xcite that i16293b drives the large e w outflow . our observations also show that the submillimeter dust emission peaks on i16293a and more or less follows the c@xmath82o emission mapped by @xcite . single dish observations have insufficient resolution to show any contribution from i16293b at all , even though we know from vla and 3 mm aperture synthesis studies ( e.g. , @xcite ) that the spectral index of i16293b is much steeper than that of i16293a and indicative of dust emission . what is i16293b ? based on the scarce evidence that we have , i16293b looks like a young t tauri star , which has driven an extremely powerful large outflow which extends @xmath286 pc on the blue - shifted side . such outflows from low - luminosity young stars are not uncommon ; they are seen both as large scale optical jets @xcite and large scale co outflows @xcite . i16293b is a low - luminosity star , @xmath287 l@xmath107 , associated with compact dust emission ; the true luminosity may be found by means of high resolution millimeter - wave observations . we derive a bolometric luminosity @xmath288 l@xmath107 for the prestellar core i16293e , and find that its fir continuum spectrum resembles that of the class 0 source i16293a ( fig . @xcite quote a dust temperature @xmath289 k for this source , higher than our value of @xmath290 k since we correct for the overlapping envelope of i16293a at the position of i16293e . this correction is important since both cores are connected with a bridge - like structure ( figs . [ f1 ] , [ f5 ] ) . the temperature and density structure of i16293e resembles the conditions in the outer envelope of i16293a . it is very interesting that the hdo abundance of the warm envelope of the yso , and the core of i16293e are similar . this suggests that some of the gas species in the warm envelope may still have the abundance that they had in the prestellar stage , the difference being that the gas in the warm envelope is excited and thus easier to detect . a comparison between i16293e and i16293a could therefore be very important to understand the formation of a first hydrostatic core , a so - called class @xmath1i protostar @xcite . such protostars lie between the prestellar clouds and the class 0 protostars , and occur when a collapsing cloud first becomes optically thick , heats up and reaches a quasi continuum stage . this stage is short lived , about @xmath291 yrs , after which a second collapse takes place and the final protostar is formed . class @xmath1i protostars are characterized by a central core of about 200 k , embedded in a cold core of about 10 k , although the models of @xcite mostly reached a central source with a much lower temperature of 20 k. is a first protostar present in i16293e ? most of the observed molecular line profiles from i16293e show two peaks with a stronger redshifted peak , characteristic of expansion motions . in [ ew ] we argue that this may be due to projection effects of pieces of the e w ouflow from i16293b which streams around i16293e . we also do not find an outflow associated with i16293e in our co maps ( [ ew ] ) , so it should be younger than a class 0 object . a central class @xmath1i protostar with @xmath292200 k may be present but very hard to detect in deep mid- or far infrared continuum imaging since the spectral signature will be only a weak shoulder at the wien side of the energy spectrum of the core . such a first protostar will be very compact and have little mass , @xmath293 , which can easily be hidden in the dense cold i16293e core . detection with scuba seems impossible because of the beam size and because the outer envelope of such a cold core will be warmer due to the interstellar radiation field and will dominate the continuum emission . our hdo emission , which we think comes from the heart of the core , has a narrow linewidth ( @xmath294 km s@xmath10 ) and our models yield a kinetic temperature of about 16 k. this indicates that a first protostar must still be rather cold at the low end of the temperature interval of @xcite . sensitive high resolution submillimeter interferometry of the continuum with , e.g. , the atacama large millimeter array may reveal a cold accretion disk . we think that ldn 1689n is a remarkable low - mass star - forming region , but not because it ` harbors a young protostar , two molecular outflows and a cold nearby molecular core ' @xcite . multiple outflows are common around ysos and have been reported , e.g. , in l1448 @xcite , hh 24 @xcite , and ngc 1333 @xcite . deep studies of embedded ysos in low - mass cores revealed the presence of apparently starless cold companion cores with masses of about a few @xmath158 @xcite . strong deuteration is expected in cold dense prestellar cores which are on the verge of forming a first hydrostatic core because of the extreme depletion of molecules like co which may result in an abundace ratio [ h@xmath4d@xmath2]/[h@xmath46]@xmath295 and a boost of the [ d]/[h ] ratio in molecules ( [ s : intro ] ) . indeed , several prestellar clouds have now been found to have a large depletion of co ( e.g. , l1544 , @xcite ; l68 , @xcite ) and a large h@xmath4d@xmath2 abundance @xcite . but , what really is remarkable in ldn 1689n is that it can be considered as a ` deuteration candle ' with the strongest deuteration known so - far in low - mass ysos . what causes the extreme deuteration in the envelope of i16923a and the core of i16293e ? the answer may lie in the fact that i16293ab is a binary system where the ysos and their outflows differ in age . i16293b is probably the oldest of the two ysos and the source of the fossilized e w outflow ( [ ew ] ) . this outflow is responsible for the dynamical interaction of a part of the eastern blue lobe with a dense dark core , the precursor of i16293e . this interaction caused a shock with velocity @xmath29610 km s@xmath10 @xcite as measured at the deuterium peak d north of i16293e and possibly higher at the first place of impact west of the core . this could have released deuterated species which were condensed on the dust grains , and compressed the gas , which then cools efficiently to low temperature @xcite . since the e w outflow causes a less energetic shock than the ne sw outflow @xcite and sio emission is found only associated with the ne sw outflow ( e.g. , @xcite ) , we expect only deuterated species to desorb since these were frozen out relatively late and thus are located near the outer surface of the grain mantles . these slow shocks of the outflow leave the grains mostly intact , with a post shock composition like that of a grain in a @xmath292 k gas . the release of deuterated species , combined with the gas - phase deuterium chemistry at low temperatures , boosts the deuterium enrichment to extremely high values just after the passage of a shock ; these deuterated species freeze out again in about @xmath297 yrs . compression from this shock resulted in an elongated n s structure , visible in the 450 @xmath0 m image ( fig . [ f5 ] ) . at the north side this filamentary structure bends to the northeast where the gentle flow passes around the core ; here the deuterium peak is located . this shock would cause a blue - shift of the systemic velocity in i16293e with respect to the parental ldn 1689n cloud . indeed , the hdo spectrum toward i16293e is blue - shifted by about 0.3 km s@xmath10 with respect to the systemic velocity of the ldn 1689n envelope toward i16293a ( table [ t1 ] ) . @xcite also found such a blue - shift of deuterated molecules toward the deuterium peak d which is located about 10 arcsec northwest of the continuum peak of i16293e . we believe that at this position the shock has the greatest chemical impact . this strong deuterated core may have been shocked again at a later time by the ne sw outflow from i16293a causing a velocity red shift of 0.3 km s@xmath10 . however , only the non - deuterated species , which predominently reside in the outer core region , are red - shifted with respect to their deuterated forms found mainly in the heart of the core . this is exactly the velocity shift which @xcite would expect from an interaction of the northeast red lobe . however , they did not detect this , since they measured only the h@xmath3co@xmath2 10 line at the position of i16293e which has a velocity equal to the cloud systemic velocity . under our hypothesis this has to be measured with respect to the reference frame of the first shock . this second shock boosted the gas - phase deuterium chemistry to its maximum and the temperature is further lowered due to further compression . we think that it is this firecracker effect which leads to the extreme deuteration of ldn 1689n in general , and i16293e in particular . multiple shocks may be common in low - mass ysos ; @xcite suggest that serpens smm1 is influenced by multiple low - velocity shocks driven by the outflow . we have analyzed our submillimeter dust and continuum data in terms of an excitatiom analysis using a model for the temperature and density structure of the disk , envelope , and outflow of i16293a , b , and e. the spectral lines of h@xmath4o , hdo , and h@xmath4d@xmath2 were subsequently compared with predictions from a simple chemical network . figure [ f17 ] shows a schematic overview of the iras 16293 - 2422 region showing its distinct physical regions ( disk , warm dense core , warm envelope , cold envelope , outflow regions , etc . ) . for each region we indicate the derived physical parameters , and their molecular abundances . for the submillimeter disk of i16293a we derive a dust temperture @xmath298 k , a source size of about 800 au , a total mass of 1.8m@xmath107 and a luminosity of 16.5 l@xmath107 which corresponds to a gas density @xmath108 @xmath30 . the envelope of i16293a can be well described by a collapsing envelope model with parameters age @xmath170@xmath299 yr and sound speed @xmath155 km s@xmath10 . in fact the center of the expansion wave , @xmath300 , is 8903700 au from the centre which corresponds to 623@xmath37 at the distance of 160 pc , comparable to the beam size of the scuba continuum and submillimeter line emission of the species we used in the analysis . the outer radius of the envelope is about 7300 au and the temparature ranges from 115 k in the inner region to 12 k in the outer region . the envelope mass is @xmath301 m@xmath107 . we find that for a yso such as i16293a the line emission is beter suited than the continuum emission to constrain the infall parameters . the excitation analysis of the envelope yields a constant hdo abundance of @xmath190 throuhout the envelope , and an ortho - h@xmath4o abundance of @xmath18 for the regions with @xmath261 k , while the abundance is lowered by a factor of 150 in the colder regions . from the c@xmath131o analysis we find that co is depleted by a facot of 50 for the envelope regions where @xmath302 k. the line wings of the co , hdo , and h@xmath4o emission can qualitatively be explained by a single outflow model with a turbulent velocity of 2.3 km s@xmath10 and where the maximum outflow velocity is about 6 km s@xmath10 . the excitation analysis of the outflow model indicates that the ortho - h@xmath4o abundance in the outflow is about @xmath303 and that the hdo abundance in the outflow is in the range of @xmath223 to @xmath172 . a more quantitative comparison requires a multiple outflow model . we find that only two outflows are present , a ne sw outflow with a dynamical time scale of about @xmath304@xmath305 yr which is powered by i16293a a class 0 source , and an e w outflow which is a fossil flow with a dynamical time scale of @xmath306@xmath307 yr driven by i16293b which looks like a young low - luminosity t tauri star . the prestellar object i16293e is well described with an isothermal core of @xmath308 k , a core radius of 8000 au and a power law density structure of the form @xmath110 where @xmath309 au@xmath310 @xmath30 . the total mass mass of this core is @xmath122 m@xmath107 . our excitation analysis of the core indicate a depletion of co and hco@xmath2 by a factor of ten relative to the ` standard abundances ' and that the dco@xmath2 abundance is larger than that of h@xmath3co@xmath2 . our tentative detection of hdo indicates the presence of a cold condensation in the heart of this core where the hdo abundance is about @xmath311 . our detailed co mapping shows that i16293e has no outflow , so its evolutionary stage is younger than class 0 . we argue that this core may hide a first hydrostatic core , a so called class @xmath1i object which is deeply embedded in a largely unaffected cold core . the reason for the extreme deuteration in the l16293n cloud and in particular in i16293a , and i16293e may lie in the fact that i16293ab is a binary system where the ysos and the outflows differ in age causing a firecracker efffect of the deuteration . in this scenario the older e w ouflow from i16293b is responsible for the formation of i16293e out of a dense dark core through a slow shock where only deuterated species were desorbed from the grains which , in combination with the low temperatures , amplifies the deuterium enrichment in the gas phase before a possible re - absorption on grains . this strong deuterated core may have been shocked again an a later time by the younger nw sw outflow from i16293a further boosting the gas - phase deuterium chemistry to extreme values . it is remarkable that the abundances of many molecules in the prestellar core i16293e and in the warm part of the envelope of i16293a are comparable . in particular , the hdo abundances are similar which indicates that the warm gas in i16293a may have gone through the same cold pre - collapse phase that i16293e is currently in . a simple chemical gas - phase modeling of the deuterium chemistry requires that co is depleted by a facor of 50 for @xmath312 k in order to reproduce the observed h@xmath4d@xmath2 emission from our chemical network applied to the envelope @xmath313 structure , and that an h@xmath4o abundance of @xmath314 for @xmath261 k and @xmath315 for @xmath17 k is required to reproduce the observed spectrum from the chemical model . the [ hdo]/[h@xmath4o ] abundance in the warm inner - envelope of i16293a ( @xmath316 ) is close to the value that has been measured in the solar system and supports this . our chemical modeling indicates that the [ hdo]/[h@xmath4o ] ratio is determined in the cold gas phase prior to star formation , and that this ratio is conserved in the early stages of low - mass star through freeze out on dust grains formation and that the molecules are released unprocessed to the gas phase after the formation of a low - mass yso . if this warm deuterated gas is subsequently locked - up in a proto - planetary disk before the high temperature gas chemistry causes a de - deuteration , the [ d]/[h ] ratio of hdo in low - mass pre - solar systems would remain conserved . to investigate this scenario in detail , sensitive follow - up observations of the hdo and h@xmath4o ground - state transitions in a large sample of low - mass ysos are needed at high spatial and spectral resolution ( i.e. , interferometry from high altitudes [ hdo ] and from space [ h@xmath4o ] ) and allow a study of the [ hdo]/[h@xmath4o ] abundance evolution from prestellar cores to planets . the h@xmath317o ground - state transition may be more suitable than the ortho - h@xmath4o transition since the ground - state of the latter molecule is at @xmath318 k above ground , while that of h@xmath317o is at 0 k , and the main isotopic form of water is expected to be mostly optically thick in its ortho and para ground - state lines . the james clerk maxwell telescope is operated by the joint astronomy centre , on behalf of the particle physics and astronomy research council of the united kingdom , the netherlands organisation for scientific research , and the national research council of canada . it is a pleasure to thank the jcmt staff and the mpifr division for submillimeter technology for their outstanding support . the canadian astronomy data center is operated by the dominian astrophysical observatory for the national research council of canada s herzberg institute of astrophysics . aikawa , y. , & herbst e. 1999 , , 526 , 314 aikawa , y. , ohashi , n. , inutsuka , s - i . , herbst , e. , & takakuwa , s. 2001 , , 552 , 639 aikawa , y. , van zadelhoff , g .- j . , van dishoeck , e. f. , & herbst , e. 2002 , , 386 , 622 andr , p. , ward - thompson , d. , & barsony , m. 1993 , , 406 , 122 ashby , m. l. n. , et al . 2000 , , 539 , l115 bally , j. , & devine , d. 1997 , in herbig - haro flows and the birth of stars , iau symp . 182 , ed . b. reipurth & c. bertout ( dordrecht : kluwer ) , 29 barsony , m. , ward - thompson , d. , andr , p. , & olinger , j. 1998 , , 509 , 733 bergin , e. a. , neufeld , d. a. , & melnick , g. j. 1999 , , 510 , l145 bergin , e. a. , alves , j. , huard , t. , & lada , c. j. 2002 , , 570 , l101 bieging , j. h. , & cohen , m. 1985 , , 289 , l5 blake , g. a. , van dishoeck , e. f. , jansen , d. j. , groesbeck , t. d. , & mundy , l. g. 1994 , , 428 , 680 bockele - morvan , d. , et al . 1998 , icarus , 133 , 147 bontemps , s. , andr , p. , terebey , s. , & cabrit , s. 1996 , , 311 , 858 boss , a. p. , yorke , h. w. 1995 , , 439 , l55 brown , p. d. , & millar , t. j. 1989 , , 237 , 661 butner , h. m. , lada , e. a. , & loren , r.b . 1995 , , 448 , 207 caselli , p. , walmsley , c. m. , zucconi , a. , tafalla , m. , dore , l. , & myers , p. c. 2002 , , 565 , 344 caselli , p. , van der tak , f. f. s. , ceccarelli , c. , & bacmann , a. 2003 , , 403 , l37 castets , a. , ceccarelli , c. , loinard , l. , caux , e. , & lefloch , b. 2001 , , 375 , 40 ceccarelli , c. , hollenbach , d. j. , & tielens , a. g. g. m. 1996 , , 471 , 400 ceccarelli , c. , castets , a. , loinard , l. , caux , e. , & tielens , a. g. g. m. 2000 , , 331 , 372 choi , m. , panis , j .- f . , & evans ii , n. j. 1999 , , 122 , 519 doty , s. d. , & neufeld , d. a. 1997 , , 489 , 122 duncan , w. d. , robson , e. i. , ade , p. a. r. , griffin , m. j. , & sandell , g. 1990 , , 243 , 126 eberhardt , p. , reber , m. , krankowsky , d. , & hodges , r. r. 1995 , , 302 , 301 eislffel , j. , & mundt , r. 1997 , , 114 , 280 emerson , d. t. , klein , u. , & haslam , c. g. t. 1979 , , 76 , 92 estalella , r. , anglada , g. , rodrguez , l. f. , & garay , g. 1991 , , 371 , 626 garay , g. , mardones , d. , rodrguez , l.f . , caselli , p. , & bourke , t. l. 2002 , , 567 , 980 gensheimer , p. d. , mauersberger , r. , & wilson , t. l. 1996 , , 314 , 281 green , s. 1989 , , 70 , 813 gulin , m. , langer , w. , & wilson , r. w. 1982 , , 107 , 107 haslam , c. g. t. 1974 , , 15 , 333 helmich , f. p. , et al . 1996a , , 315 , l173 helmich , f. p. , van dishoeck , e. f. , & jansen , d. j. 1996b , , 313 , 657 herbst , e. 1982 , , 252 , 810 hildebrand , r. h. 1983 , , 24 , 267 hirano , n. , mikami , h. , umemoto , t. , yamamoto , s. , taniguchi , y. 2001 , , 547 , 899 hogerheijde , m. r. & sandell , g. 2000 , , 534 , 880 hogerheijde , m. r. & van der tak , f. f. s. 2000 , , 362 , 697 holland , w. s. , gear , w. k. , lightfoot , j. f. , jenness , t. , robson , e. i. , cunningham , & c. r. , laidlaw , k. 1998 , in proc . spie vol . 3357 , advanced technology mmw , radio , and terahertz telescopes , ed . t. g. phillips ( bellingham : spie press ) , 305 jacq , t. , walmsley , c. m. , henkel , c. , baudry , a. , mauersberger , r. , & jewell , p. r. 1990 , , , 228 , 447 jenness , t. & lightfoot , j. f. 1998 , in asp conf . ser . 145 , astronomical data analysis software and systemes vii , ed . r. albrecht et al . ( san francisco : asp ) , 216 jensen , m. j. , bilodeau , r. c. , safvan , c. p. , seiersen , k. , andersen , l. h. , pedersen , h. b. , & heber , o. 2000 , , 534 , 764 knee , l. b. g. & sandell , g. 2000 , , 361 , 671 lada , c. j. , 1991 , in nato - asi ser . 342 , the physics of star formation and early stellar evolution , . ed . c. j. lada & n. d. kylafis ( dordrecht : kluwer ) , 329 larsson , b. , liseau , r. , & menshchikov , a. b. 2002 , , 386 , 1055 lcuyer , c. , gillet , ph . , & robert , f. 1998 , chemical geology , 145 , 249 lepp , s. , dalgarno , a. , & sternberg , a. 1987 , , 321 , 383 lis , d. c. , grin , m. , phillips , t. g. , & motte , f. 2002 , , 569 , 322 liseau , r. , & sandell , g. 1986 , , 304 , 459 looney , l. w. , mundy , l. g. , & welch , w. j. 2000 , , 529 , 477 meier , r. , owen , t. c. , matthews , h. e. , jewitt , d. c. , bockele - morvan , d. , biver , n. , crovisier , j. , & gauthier , a. 1998 , science , 279 , 842 melnick , g. j. , stauffer , j. r. , ashby , m. l. n. , bergin , e. a. , chin , g. 2000 , , 539 , l77 mezger , p. g. , sievers , a. , zylka , r. , haslam , c. g. t. , kreysa , e. , & lemke , r. 1992 , , 265 , 743 millar , t. j. , bennett , a. , & herbst , e. 1989 , , 340 , 906 millar , t. j. , herbst , e. , & charnley , s. b. 1991 , , 369 , 147 mizuno , a. , fukui , y. , nozawa , s. , & takano , t. 1990 , , 356 , 184 motte , f. , andr , p. , & neri , r. 1998 , , 336 , 150 mundy , l. g. , wilking , b. a. , & myers , s. t. 1986 , , 311 , l75 mundy , l. g. , wootten , a. , & wilking , b. a. 1990 , , 352 , 159 mundy , l. g. , wootten , a. , wilking , b. a. , blake , g. a. , & sargent , a. i. 1992 , , 385 , 306 narayanan , g. , walker , c. k. , & buckley , h. d. 1998 , , 496 , 292 neufeld , d. a. , ashby m. l. n. , bergin , e. a. , chin , g. , erickson , n. r. 2000 , , 539 , l107 olberg , m. , bester , m. , rau , g. , et al . 1985 , , 142 , l1 ossenkopf , v. & henning , th . 1994 , , 291 , 943 padman , r. , bence , s. , & richer , j. 1997 , in herbig - haro flows and the birth of stars , iau symp . 182 , ed . b. reipurth & c. bertout ( dordrecht : kluwer ) , 123 phillips , t. g. , van dishoeck , e. f. , & keene , j. 1992 , , 399 , 533 phillips , t. r. , maluendes , s. , & green , s. 1996 , , 107 , 467 pineau des forts , g. , roueff , e. , & flower , d. r. 1989 , , 240 , 167 richer , j. s. 1992 , , 254 , 165 rodgers , s. d. , & millar , t. j. , 1996 , 280 , 1046 rodrguez , l. f. , cant , j. , torelles , j. m. , & ho , p. t. p. 1986 , , 301 , l25 sandell , g. 1994 , , 271 , 75 sandell , g. , jessop , n. , and jenness , t. 2001 , the scuba map reduction cookbook , starlink cookbook 11.2 , http://www.starlink.rl.ac.uk/star/docs/sc11.htx/sc11.html#xref sault , r. j. , teuben , p. j. , & wright , m. c. h. , 1995 , in asp conf . 77 , astronomical data analysis software and systemes iv , ed . r. a. shaw et al . ( san francisco : asp ) , 433 schier , f. l. , jrgensen , j. k. , van dishoeck , e. f. , & blake , g. a. 2002 , , 390 , 1001 schier , f. l. , jrgensen , j. k. , van dishoeck , e. f. , & blake , g. a. 2003 , in chemistry as a diagnostic of star formation , ed . c. l. curry & m. fich ( ottawa : nrc press ) , in press shah , r. y. & wootten a. 2001 , , 554 , 933 shirley , y. l. , evans , n. j. , ii , rawlings , j. m. c. , & gregersen , e. m. 2002 , , 131 , 249 shu , f. h. 1977 , , 214 , 488 schulz , a. , gsten , r. , walmsley , c. m. , & serabyn , e. 1991 , , 246 , l55 smith , d. , adams , n.g . , & alge , e. 1982 , , 263 , 123 snell , r. l. , howe , j. e. , ashby , m. l. n. , bergin , e. a. , chin , g. , et al . 2000 , , 539 , l101 stark , r. , van der tak , f. s. s. , & van dishoeck , e. f. 1999 , , 521 , l67 tielens , a. g. g. m. 1989 , in proc . iau symp . 135 , interstellar dust , ed . l. j. allamandola & a. g. g. m. tielens ( dordrecht : kluwer ) , 239 van dishoeck , e. f. , blake , g. a. , jansen , d. j. , & groesbeck , t. d. 1995 , , 447 , 760 walker , c. k. , lada , c. j. , young , e. t. , maloney , p. r , & wilking , b. a. 1986 , , 309 , l47 walker , c. k. , lada , c. j. , young , e. t. , & margulis , m. 1998 , , 332 , 335 walker , c. k. , carlstrom , j. e. , & bieging , j. h. 1993 , , 402 , 655 watson , w. d. , anichich , v. g. , & huntress , w. t. , jr . 1976 , , 205 , l165 williams , j. p. , bergin , e. a , caselli , p. , myers , p. c. , & plume , r. 1998 , , 503 , 689 wilson , t. l. & rood , r. 1994 , , 32 , 191 wolf - chase , g. a. , barsony , m. , & olinger , j. 2000 , , 120 , 1467 wootten , a. , & loren , r. b. 1987 , , 317 , 220 wootten , a. 1987 , in proc . iau symp . 120 , astrochemistry , ed . m. s. vardya & s. p. tarafdar ( dodrecht : kluwer ) , 311 wootten , a. 1989 , , 337 , 858 lrrr 2.0mm/27@xmath37 ( 2 ) & @xmath319 & + 1.3mm/19.5@xmath37 ( 1 ) & @xmath320 & @xmath321 & + 1.1mm/18.5@xmath37 ( 3 ) & @xmath322 & @xmath323 & + 850/14.5@xmath37 & & @xmath324 & 16.0 + 850/14.0@xmath37 & @xmath325 & @xmath326 + 800/16.5@xmath37 ( 2 ) & @xmath327 & @xmath328 & + 750/13.3@xmath37 & & @xmath329 & + 450/18@xmath37 ( 2 ) & @xmath330 & @xmath331 & 102.7 + 450/8.0@xmath37 & @xmath332 & @xmath333 + 350/19@xmath37 ( 1 ) & @xmath334 & @xmath335 & + lcrrrr + c@xmath82o & 32 & 33.3 & & & 4.0 + & 21 & 18.4 & 6.50 & 2.50 & 3.76 + c@xmath131o & 21 & 9.3 & 3.24 & 2.86 & 3.99 + @xmath61co & 76 & 473 & & & 3.9 + & 65 & 408 & & & 3.9 + hco@xmath2 & 43 & 81 & & & 4.1 + & 32 & 70 & & & 4.0 + h@xmath3co@xmath2 & 43 & 7.2 & & & 4.7 + h@xmath3co@xmath2 & 32 & 11.8 & 4.87 & 2.27 & 4.23 + & 32 & @xmath336 & @xmath337 & 0.78 & 4.24 + dco@xmath2 & 54 & 3.1 & & & 4.0 + & 32 & 5.18 & 2.56 & 2.02 & 4.48 + h@xmath4d@xmath2 & @xmath7@xmath8 & 0.8 & 1.1 & 0.6 & 3.9 + n@xmath4h@xmath2 & 43 & 4.4 & 2.6 & 1.7 & 3.7 + hdo & @xmath5@xmath6 & 6.0 & 1.0 & 5.9 & 4.3 + & @xmath5@xmath6 & @xmath338 & @xmath339 & 0.60 & 4.2 + h@xmath4o & @xmath7@xmath5 & 2.9 & & & 3.9 + + c@xmath82o & 21 & 5.40 & 3.60 & 1.41 & 3.80 + c@xmath131o & 21 & 1.70 & 1.26 & 1.27 & 4.03 + h@xmath3co@xmath2 & 32 & 1.64 & 1.25 & 1.23 & 4.08 + dco@xmath2 & 32 & 3.74 & 3.14 & 1.12 & 3.76 + hdo & @xmath7@xmath8 & 0.1 & 0.15 & 0.6 & 3.5 + lrrrrr c@xmath82o & @xmath340 & 0.83 & @xmath341 & @xmath342 & + c@xmath131o & @xmath343 & 0.76 & @xmath344 & @xmath345 & + hco@xmath2 & @xmath346 & 0.69 & @xmath347 & @xmath348 & + h@xmath3co@xmath2 & @xmath349 & 0.50 & @xmath349 & @xmath350 & + dco@xmath2 & @xmath351 & 0.67 & @xmath349 & @xmath352 & + h@xmath4o & & & @xmath353 & & @xmath354 + & & & @xmath355 & & + hdo & @xmath356 & 0.20 & @xmath357 & & + h@xmath4d@xmath2 & & & @xmath358 & & + & & & @xmath359 & & + n@xmath4h@xmath2 & & & @xmath360 & & + h@xmath38o@xmath2 & & & @xmath361 & & + lrrrr dynamical time scale ( yr ) & @xmath362 & @xmath363 & @xmath364 & @xmath365 + mass ( m@xmath13 ) & 0.10 & 0.05 & 0.33 & 0.18 + momentum ( m@xmath13 km s@xmath10 ) & 0.60 & 0.22 & 2.27 & 1.25 + energy ( m@xmath13 km@xmath366 s@xmath367 ) & 3.44 & 0.78 & 23.4 & 8.64 + force ( 10@xmath368 m@xmath13 km s@xmath10 yr@xmath10 ) & 1.91 & 0.74 & 6.43 & 4.90 +	we present deep images of dust continuum emission at 450 , 800 , and 850 @xmath0 m of the dark cloud ldn 1689n which harbors the low - mass young stellar objects ( ysos ) iras 16293@xmath12422a and b ( i16293a and i16293b ) and the cold prestellar object i16293e . toward the positions of i16293a and e we also obtained spectra of co - isotopomers and deep submillimeter observations of chemically related molecules with high critical densities ( hco@xmath2 , h@xmath3co@xmath2 , dco@xmath2 , h@xmath4o , hdo , h@xmath4d@xmath2 ) . to i16293a we report the detection of the hdo @xmath5@xmath6 and h@xmath4o @xmath7@xmath5 ground - state transitions as broad self - reversed emission profiles with narrow absorption , and a tentative detection of h@xmath4d@xmath2 @xmath7@xmath8 . to i16293e we detect weak emission of subthermally excited hdo @xmath5@xmath6 . based on this set of submillimeter continuum and line data we model the envelopes around i16293a and e. the density and velocity structure of i16293a is fit by an inside - out collapse model , yielding a sound speed of @xmath9=0.7 km s@xmath10 , an age of @xmath11=(0.62.5)@xmath12 yr , and a mass of 6.1 m@xmath13 . the density in the envelope of i16293e is fit by a radial power law with index @xmath14 , a mass of 4.4 m@xmath13 , and a constant temperature of 16 k. these respective models are used to study the chemistry of the envelopes of these pre- and protostellar objects . we made a large , fully sampled co @xmath15=21 map of ldn 1689n which clearly shows the two outflows from i16293a and b , and the interaction of one of the flows with i16293e . an outflow from i16293e as reported elsewhere is not confirmed . instead , we find that the motions around i16293e identified from small maps are part of a larger scale fossil flow from i16293b . modeling of the i16293a outflow shows that the broad hdo , water ground - state , and co @xmath15=65 and 76 emission lines originate in this flow , while the hdo and h@xmath4o line cores originate in the envelope . the narrow absorption feature in the ground - state water line is due to cold gas in the outer envelope . the derived h@xmath4o abundance is @xmath16 in the cold regions of the envelope of i16293a ( @xmath17 k ) , @xmath18 is warmer regions of the envelope ( @xmath19 k ) , and @xmath20 in the outflow . the hdo abundance is constant at a few @xmath21 throughout the envelopes of i16293a and e. because the derived h@xmath4o and hdo abundances in the two objects can be understood through shock chemistry in the outflow and ion - molecule chemistry in the envelopes , we argue that both objects are related in chemical evolution . the [ hdo]/[h@xmath4o ] abundance ratio in the warm inner envelope of i16293a of a few times 10@xmath22 is comparable to that measured in comets . this supports the idea that the [ hdo]/[h@xmath4o ] ratio is determined in the cold prestellar core phase and conserved throughout the formation process of low - mass stars and planets .
the standard model ( sm)-like higgs scalar with mass around 125 gev @xcite discovered at the lhc needs some mechanism for stabilizing it against high - energy scale quantum correction . among the well known candidates which achieve this naturally low - scale supersymmetry ( susy ) is expected to play an important role at the tev scale . however , the first run of lhc has not observed any signal of new physics yet , and pushes the susy particle masses into multi - tevs region @xcite . so the absence of susy particles near the weak scale @xmath0 , together with the observed higgs mass , severely challenge the low - scale susy . in the context of minimal supersymmetric standard model ( mssm ) stop masses far above the weak scale is required by the observed higgs mass when the mixing effect is weak . given susy mass spectrum far above the weak scale , the naturalness is spoiled naively . however , this statement can be relaxed in some specific situations . susy with focusing phenomenon @xcite , which is named as focus point susy , is few of such examples . in focus point susy the sensitivity of up - type higgs mass squared to the mass scale of susy mass spectrum is suppressed because of cancellation among the large renormalization group ( rg ) corrections . this phenomenon leads to a dramatical reduction of fine tuning associated with electroweak symmetry breaking ( ewsb ) . as a result , it provides us an alternative choice of natural susy consistent with the lhc data and the observed higgs mass . unfortunately , focus point susy can not be realized in the minimal setup from the viewpoint of model building such as conventional supergravity @xcite and the minimal gauge mediation ( gm ) @xcite . however , they are expected to be achieved in some subtle cases . for recent examples in the context of gm , see , e.g. , @xcite . these examples are restricted to special susy - breaking mediation scale @xmath1 . based on our earlier work @xcite , we will take an arbitrary @xmath1 and generalize results above to general focus point . consider the fact that the cold dark matter demands some electroweakino mass around the weak scale , we will focus on the high - scale susy in which gaugino masses are light in compared with other susy particle masses . the paper is organized as follows . firstly , we determine the conditions for focus point in susy with arbitrary @xmath1 in section ii . then , we analyze the prediction for the higgs mass in focus point susy in section iii . we find that the observed higgs mass requires the input value for @xmath2 of order @xmath3 multi - tevs . with such magnitude of @xmath2 the observed higgs mass can be explained in a wide classes of high - scale susy models . the second part of this paper is devoted to the realization of focus point in high - scale susy . in section iv we will construct concrete and complete examples by employing non - minimal gm . we will show that for the case of small @xmath4 term focus point susy is viable for a wide range of @xmath1 , which generalizes previous results in the literature . for the case of large @xmath4 term , we find that focus point susy is viable in a large classes of gm with direct yukawa coupling between the messengers and the mssm singlet . finally we discuss our results in section v. the calculation of soft masses is included in appendix a. we begin with the conditions for focus point in susy . given light gauginos , in compared with soft masses @xmath5 , @xmath6 , @xmath2 and @xmath4 term squared , the one - loop renormalization group equations ( rges ) for them are simply given by , @xmath7 where @xmath8 , with @xmath9 the rg running scale and @xmath1 the susy - breaking mediation scale . here @xmath10 denotes the top yukawa coupling . all other sm yukawa couplings will be ignored in our discussion . the correlation for these soft masses between scale @xmath1 and the weak scale @xmath11 gev is obtained in terms of solving the rges in eq .. in particular , we have @xmath12\left(m^{2}_{\tilde{q}_{3}}[m]+ m^{2}_{\tilde{u}_{3}}[m ] + m^{2}_{h_{u}}[m]-a^{2}_{t}[m]\right)\nonumber\\ & + & i[m]^{2}a^{2}_{t}[m]-\left(m^{2}_{\tilde{q}_{3}}[m]+ m^{2}_{\tilde{u}_{3}}[m]-m^{2}_{h_{u}}[m]\right),\end{aligned}\ ] ] where the real coefficient @xmath13 $ ] is defined as , @xmath14=\exp\left ( 6\times \int^{\ln\frac{v}{m}}\frac { y_{t}^{2}[t']}{8\pi^{2 } } dt ' \right).\end{aligned}\ ] ] for the case with negligble @xmath15 , eq . reduces to the well known result @xcite , @xmath16\left(m^{2}_{\tilde{q}_{3}}[m]+ m^{2}_{\tilde{u}_{3}}[m ] + m^{2}_{h_{u}}[m]\right)-\left(m^{2}_{\tilde{q}_{3}}[m]+ m^{2}_{\tilde{u}_{3}}[m]-m^{2}_{h_{u}}[m]\right).\end{aligned}\ ] ] as well known , for soft mass spectrum far above the weak scale in eq . there is a large fine tuning naively . but this can be avoided if there is significant cancellation among large contributions to up - higgs soft mass squared in eq .. generally speaking , this cancellation does not happen in the minimal setup of model building . however , if it indeed happens in such as focus point susy , with the cancellation referred as focusing phenomenon , one can derive the focus condition , regarding eq .. for @xmath17 gev . any point in each focusing line generates the focusing phenomenon , @xmath18\simeq 0 $ ] . see the text for comments on references.,width=432 ] in fig.[focuscondition ] we show the `` focusing '' line in the plane of @xmath20 and @xmath21 for @xmath17 gev . each point on the line gives rise to the focusing phenomenon . earlier works in @xcite discussed focusing in gut - scale model without @xmath27 term . in these cases @xmath28\simeq 1/3 $ ] , so they correspond to the focus point @xmath29 in fig.[focuscondition ] . work in @xcite addressed similar situation but with sizable @xmath27 term , it corresponds to the gray line . work in @xcite discussed intermediate scale model with @xmath30 gev , which corresponds to the blue line . in this section we discuss the prediction for the higgs mass in high - scale susy with focus point . in particular , we would like to show which points in the focus lines in fig.[focuscondition ] can explain the observed higgs mass @xmath31 gev @xcite at the lhc . it is well known that the tree - level contribution to the higgs mass is up bounded by the @xmath32 boson mass , so loop correction must be taken into account for the higgs mass fit . moreover , contributions to the higgs mass higher than the two - loop order should be considered when the squark masses are larger than @xmath33 tev . it has been shown in @xcite that the three - loop contribution gives rise to @xmath34 gev uplifting of the higgs mass . for this order approximation one can either use the numerical program in @xcite , or as we choose in this paper follow the three - loop analytic formula for the higgs mass in ref.@xcite . for @xmath44 and @xmath17 gev . the focusing lines of fig.[focuscondition ] are shown simultaneously . in each panel , we show the sensitivity of higgs mass to the input mass parameter @xmath45 . it clearly shows that the observed higgs mass constrains @xmath45 in the range @xmath46 tev @xmath47 tev.,title="fig:",width=268 ] for @xmath44 and @xmath17 gev . the focusing lines of fig.[focuscondition ] are shown simultaneously . in each panel , we show the sensitivity of higgs mass to the input mass parameter @xmath45 . it clearly shows that the observed higgs mass constrains @xmath45 in the range @xmath46 tev @xmath47 tev.,title="fig:",width=268 ] for @xmath44 and @xmath17 gev . the focusing lines of fig.[focuscondition ] are shown simultaneously . in each panel , we show the sensitivity of higgs mass to the input mass parameter @xmath45 . it clearly shows that the observed higgs mass constrains @xmath45 in the range @xmath46 tev @xmath47 tev.,title="fig:",width=268 ] for @xmath44 and @xmath17 gev . the focusing lines of fig.[focuscondition ] are shown simultaneously . in each panel , we show the sensitivity of higgs mass to the input mass parameter @xmath45 . it clearly shows that the observed higgs mass constrains @xmath45 in the range @xmath46 tev @xmath47 tev.,title="fig:",width=268 ] we show in fig.[higgs ] the contours of three - loop higgs mass as projected to the two - parameter plane of @xmath48 for @xmath44 and @xmath17 gev . each panel corresponds to different values of @xmath1 , for which the focusing line in fig.[focuscondition ] is shown simultaneously . it is shown that @xmath49\simeq 3 $ ] tev for small mixing effect ( @xmath50 ) , which is consistent with the three - loop result in ref . @xcite for the case of degenerate squark masses . in each panel , one observes the sensitivity of higgs mass to the input mass parameter @xmath45 , which clearly indicates that the observed higgs mass constrains @xmath45 in the range @xmath51 tev @xmath52 tev . this range for @xmath45 is subject to the choice on parameter @xmath53 , as @xmath53 determines the tree - level contribution . we show the correlation between the range and @xmath53 in fig.[higgss ] , which suggests that for @xmath54 we have @xmath55 tev @xmath56 tev instead . if one takes smaller @xmath53 , it is expected that larger @xmath57 is needed . for @xmath54 . in this case there is about @xmath58 gev reduction in the tree - level part of higgs mass in compared with the previous choice @xmath44 in fig.[higgs ] , so larger loop contribution to the higgs mass is required.,title="fig:",width=268 ] for @xmath54 . in this case there is about @xmath58 gev reduction in the tree - level part of higgs mass in compared with the previous choice @xmath44 in fig.[higgs ] , so larger loop contribution to the higgs mass is required.,title="fig:",width=268 ] for @xmath54 . in this case there is about @xmath58 gev reduction in the tree - level part of higgs mass in compared with the previous choice @xmath44 in fig.[higgs ] , so larger loop contribution to the higgs mass is required.,title="fig:",width=268 ] for @xmath54 . in this case there is about @xmath58 gev reduction in the tree - level part of higgs mass in compared with the previous choice @xmath44 in fig.[higgs ] , so larger loop contribution to the higgs mass is required.,title="fig:",width=268 ] here a few comments are in order , regarding the fit to the observed higgs mass in fig.[higgs ] and fig.[higgss ] . @xmath59 , given the same @xmath45 , compare the four contours of higgs mass in the four panels in either fig.[higgs ] or fig.[higgss ] . it is shown that the focus point value of @xmath21 at the crossing point decreases as @xmath1 approaches to the gut scale . the reason is partially due to the fact that @xmath13 $ ] decreases as @xmath1 increases . @xmath60 , given the same @xmath1 , compare the contours of higgs mass in each panel in either fig.[higgs ] or fig.[higgss ] . as expected , the focus point value of @xmath21 ( @xmath20 ) at the crossing point increases ( decreases ) as @xmath45 increases . moreover , there are more crossing points in the case with larger @xmath1 , in this sense the focusing may be more easily achieved in high - scale other than low - scale susy . @xmath61 , for @xmath57 with such order of magnitude , the higgs couplings of higgs sector are similar to those in the decoupling region . the average stop mass is above @xmath33 tev for arbitrary @xmath1 , which has no conflict with the lhc 2013 data . so it is probably impossible to probe so heavy stops at the second run of lhc , and light electroweakinos will be the smoking gun of such high - scale susy with focus point . in summary , the combination of fig . [ higgs ] and fig . [ higgss ] implies that given a focusing line referring to @xmath1 most of focus points are consistent with the observed higgs mass at the lhc by adjusting the underlying and overall energy scale @xmath45 in the soft mass spectrum . typically , @xmath57 is in the range of @xmath62 $ ] tev and @xmath63 $ ] tev for @xmath44 and @xmath64 , respectively . in the next section , we will proceed to construct high - scale susy with focus point which can explain the observed higgs mass . in this section we consider the realization of focusing phenomenon in high - scale susy . we restrict us to gm @xcite for this purpose . firstly we discuss gauge mediated focus point in high - scale susy with small @xmath4 term . in the context of gm , there are two important observations for our discussion . @xmath65 , the boundary value @xmath66 $ ] disappears at the one - loop level when there are no direct yukawa - like couplings between the messenger field(s ) and mssm matter field(s ) in the superpotential . @xmath67 , the gaugino masses vanish at the same order when the mass matrix for messenger fields @xmath68 satisfies @xmath69 . @xmath70 , the scalar soft masses do not disappear at this order generally . as shown in @xcite , points @xmath65-@xmath70 can be satisfied in a simple model . in this model the messenger fields are a set of chiral and bi - fundamental supermultiplets @xmath71 , @xmath72 , @xmath73 , @xmath74 that transforme under @xmath75 as , @xmath76 the renormalizable superpotential consistent with sm gauge symmetry in this model reads as , @xmath77 where susy - breaking sector @xmath78 , yukawa couplings @xmath79s and tree - level masses @xmath80s are assumed to be real for simplicity . it can be easily verified that both the mass matrixes @xmath68s for messenger vector @xmath81 and @xmath82 both satisfy @xmath69 . so the gaugino masses indeed vanish at the one - loop order of @xmath83 as desired . the soft scalar masses differ from those of the minimal gm , which are given by , respectively , @xmath84&= & \mathcal{a}\times \left(\frac{4}{3}s_{3}\alpha^{2}_{3}[m]+\frac{3}{4}s_{2}\alpha^{2}_{2}[m]+ \frac{1}{60}s_{1}\alpha^{2}_{1}[m]\right),\nonumber\\ m^{2}_{\tilde{u}_{3}}[m]&=&\mathcal{a}\times \left(\frac{4}{3}s_{3}\alpha^{2}_{3}[m]+ \frac{12}{45}s_{1}\alpha^{2}_{1}[m]\right),\\ m^{2}_{h_{u}}[m]&= & \mathcal{a}\times \left(\frac{3}{4}s_{2}\alpha^{2}_{2}[m]+ \frac{3}{20}s_{1}\alpha^{2}_{1}[m]\right).\nonumber\end{aligned}\ ] ] here @xmath85 , which determines the overall magnitude of soft masses above , and @xmath86 are given by @xmath87 note that the mass spectrum in eq . reduces to that of the minimal gm when @xmath88 , @xmath89&= & \mathcal{a}\times \left(\frac{4}{3}\alpha^{2}_{3}[m]+\frac{3}{4}\alpha^{2}_{2}[m]+ \frac{1}{60}\alpha^{2}_{1}[m]\right),\nonumber\\ m^{2}_{\tilde{u}_{3}}[m]&=&\mathcal{a}\times \left(\frac{4}{3}\alpha^{2}_{3}[m]+ \frac{12}{45}\alpha^{2}_{1}[m]\right),\nonumber\\ m^{2}_{h_{u}}[m]&= & \mathcal{a}\times \left(\frac{3}{4}\alpha^{2}_{2}[m]+ \frac{3}{20}\alpha^{2}_{1}[m]\right).\end{aligned}\ ] ] as function of @xmath1 for @xmath90 and @xmath91 from top to bottom line , respectively , with mssm below scale @xmath1 assumed.,width=432 ] ( solid ) and @xmath22 ( dotted ) for @xmath90 and @xmath92 @xmath93 in green ( gray ) curve.,width=432 ] now we consider focus point in high - scale susy with large @xmath4 term . in the model discussed in this subsection , the messenger fields are the same as previously studied in eq . , except that we add a singlet of sm gauge group @xmath105 and its bi - fundamental field @xmath106 into the messenger sector . these messengers are coupled to the susy - breaking spurion field @xmath78 through superpotential , @xmath107 we will generalize the number of messenger fields to @xmath108 pairs , and simply take universal couplings between messengers and @xmath109 fields and universal masses @xmath110 for our analysis . similar to our previous observations , gaugino masses vanish at the one - loop level of order @xmath83 in our setup . for our purpose we further deform the model defined in eq.([s2 ] ) by adding a direct yukawa coupling between the lepton - like messenger @xmath111 and @xmath112 , @xmath113 this new superpotential can be argued to be natural by imposing hidden parity . as mentioned in @xcite , eq . can be protected in terms of either imposing some hidden matter parity other than @xmath114 parity or global @xmath115 symmetry . if so , eq . ensures an one - loop @xmath4 term of order @xmath83 at the messenger scale . in this model scalar soft mass spectrum includes the contribution from eq.([s2 ] ) and the new one induced by yukawa coupling @xmath79 in eq .. for the case of small susy breaking ( @xmath116 ) the former is given by , @xmath117&= & \mathcal{a } ' \cdot \left(\frac{4}{3}\alpha^{2}_{3}[m]+\frac{3}{4}\alpha^{2}_{2}[m]+ \frac{1}{60}\alpha^{2}_{1}[m]\right),\nonumber\\ m^{2}_{\tilde{u}_{3}}[m]&=&\mathcal{a } ' \cdot \left(\frac{4}{3}\alpha^{2}_{3}[m]+ \frac{12}{45}\alpha^{2}_{1}[m]\right),\nonumber\\ m^{2}_{h_{u}}[m]&= & \mathcal{a } ' \cdot \left(\frac{3}{4}\alpha^{2}_{2}[m]+ \frac{3}{20}\alpha^{2}_{1}[m]\right).\end{aligned}\ ] ] with @xmath118 . while the later one reads as , @xmath119&\simeq&\frac{1}{2}\mathcal{a } ' \cdot \left(-\frac{n}{2}\alpha_{t}[m]\alpha_{\lambda}[m]\right),\nonumber\\ \delta m^{2}_{\tilde{u}_{3}}[m]&\simeq&\frac{1}{2}\mathcal{a } ' \cdot\left(-n\alpha_{t}[m]\alpha_{\lambda}[m]\right),\nonumber\\ \delta m^{2}_{h_{u}}[m]&\simeq&\frac{1}{2}\mathcal{a } ' \cdot \left\{\frac{1}{2 } n(n+3)\alpha^{2}_{\lambda}[m]-n\alpha_{\lambda}[m]\left(\frac{3}{10}\alpha_{1}[m]+\frac{3}{2}\alpha_{2}[m]\right)\right\},\nonumber\\ \delta a^{2}_{t}[m]&\simeq&\frac{1}{2}\mathcal{a } ' \cdot \left(\frac{1}{2}n^{2}\alpha^{2}_{\lambda}[m]\right).\end{aligned}\ ] ] here a few comments are in order , regarding eq.([softnew ] ) . the contribution to soft scalar mass spectrum due to yukawa - type interaction has been previously studied in @xcite . however , they can not be directly applied to our case . because the susy - breaking spurion superfield @xmath109 differs from the conventional situation in which @xmath120 , with @xmath121 . we present the derivation of eq.([softnew ] ) in appendix a. focusing phenomenon in this model has been discussed in @xcite for intermediate scale @xmath30 gev . here , our discussions are more general , with scale @xmath1 unfixed . for @xmath17 gev . it has been verified that each solution satisfies the consistent conditions in eq .. ,width=432 ] in this model the input parameters related to focusing conditions eq . and eq . at high energy scale @xmath1 are composed of the number of messenger pairs @xmath108 , yukawa coupling constant @xmath122 $ ] and scale @xmath1 . the other two parameters @xmath123 and @xmath53 are related to the observed higgs mass at the weak scale . note that requiring grand unification of sm gauge couplings leads to @xmath124 , and requiring the absence of landau pole leads to an upper bound on @xmath125 $ ] , although the later of which can be relaxed by either embedding the model into strong or conformal dynamics . we discuss these two parameters in the range below , @xmath126\leq 1/4\pi.\end{aligned}\ ] ] in fig.[na ] we show the solution to focusing condition eq . in the two - parameter plane of @xmath127 for different values of @xmath1 . we have verified that each solution satisfies the consistent conditions in eq .. fig.[na ] clearly indicates that focusing may happen for small @xmath128 and small @xmath122\sim 0.2 $ ] when @xmath1 is near the gut scale . conversely , large @xmath129 and large @xmath122\simeq 0.18 $ ] are required when @xmath1 is small @xmath130 gev . the rg effect on @xmath131 actually excludes such large @xmath125 $ ] in perturbative theory . for example , it requires @xmath132 gev for the representative value @xmath125= 1/4\pi$ ] . in summary , focus point high scale susy can be realized in non - minimal gm . but only large @xmath1 is allowed for the case with large @xmath4 term . in the next subsection , we address the problem of small gaugino mass in general focus point susy , which can be reconciled with the lhc 2013 data in terms of reasonable modification to the original messenger sector . the lhc 2013 data @xcite has reported gluino mass bound about @xmath133 tev . this mass bound is not satisfied in the focus point susy discussed so far . for example , the gluino mass relative to @xmath134 at scale @xmath1 in the case with large @xmath4 discussed in iv.b is given by , @xmath135}{m_{0 } } \simeq 4\sqrt{2}\pi^{2}\cdot\frac{s^{5/2}_{3}\alpha_{3}[m]}{\left(\frac{3}{4}s_{2}\alpha^{2}_{2}[m]+ \frac{3}{20}s_{1}\alpha^{2}_{1}[m]\right)^{3/2 } } \frac{m^{2}_{0}}{m^{2}}\cdot\mathcal{f}\left(\frac{-s_{3}+\sqrt{s^{2}_{3}+4}}{2},\frac{-1+\sqrt{5}}{2}\right)\nonumber\\\end{aligned}\ ] ] where function @xmath136 is defined in @xcite . together with the constraint on @xmath57 due to observed higgs mass as shown in the section iii , one can estimate the boundary value of @xmath137 $ ] . implies that the gluino mass is far below the present lower bound . some reasonable modification should be taken into account in order to complete the discussions about model building . now we re - examine the smallness of gaugino masses , which attributes to the fact that @xmath69 and consequently @xmath138 . when we employ small tree - level mass terms for some of messengers such as @xmath139 it will lead to the replacement in @xmath140 for quark and lepton messengers @xcite , @xmath141 if so , the correction to soft scalar mass spectrum and gaugino masses is of order @xmath142 and @xmath143 , respectively . provided @xmath144 the former correction can be very weak so that the focusing still holds , but the later one can be large enough to reconcile with the lhc gaugino mass bound . for example , we choose @xmath145 gev . for @xmath146 tev we have @xmath147 gev , and further @xmath148 tev from eq.([g2 ] ) . in this model both the bino and wino masses are around @xmath149 tev , so they are the main target at the 14-tev lhc . the discussions about model building in high - scale susy with focus point are therefore completed . in this paper we have explored focus point in the high - scale susy which has weak - scale electroweakino masses . we have derived conditions for the focusing phenomenon in general . we have analyzed in detail the prediction for the higgs mass in such focus point susy . the observed higgs mass at the lhc requires the input value @xmath2 of order @xmath150 tev and @xmath33 tev for @xmath151 , respectively . we also address the model building of focus point susy by employing non - minimal gm . the main results include @xmath152 for the case with small @xmath4 term , focus point is allowed for a wide range of @xmath1 ; @xmath153 for the case with large @xmath4 term , focus point is only permitted in high - scale susy . one may worry about the stability of focus point discussed so far by following two facts . one fact is that the soft scalar mass spectrum is directly related and thus sensitive to the underlying mass scales @xmath154 and @xmath1 . the other one is that the focusing phenomenon imposes constraint on these soft masses as shown in eq.([mh ] ) . so , the significant reduction due to focus point seems to be spoiled for a small change of either @xmath154 or @xmath1 . however , this is not true . because only the overall magnitude of soft mass spectrum but not their relative ratios is determined by @xmath155 . as shown in eq.([condition ] ) the focusing condition eq.([mh ] ) is actually not dependent on either @xmath154 or @xmath1 , and only dependent on @xmath1 indirectly through rg effect . once we fix @xmath1 , for example in some gut - scale susy models , the focus point induced by the accidental cancellation , is actually stable against to high mass scale . this work is supported in part by the natural science foundation of china under grant no.11247031 and 11405015 . the calculation of soft mass spectrum in gm can be performed in a few ways . among them evaluating feynman diagrams is the most direct approach @xcite . however , the number of feynman diagrams for calculating these soft scalar mass is usually large , which limits its application broadly . in @xcite a new technique is proposed , where the authors suggest analytical continuation into superspace @xmath156 in the renormalized wave function , where @xmath1 and @xmath154 refers to the messenger mass scale and the strength of susy breaking , respectively . nevertheless , this method is only valid for the standard spurion superfield @xmath157 and for the case of small susy breaking . therefore , for non - standard spurion field @xmath158 as studied in this paper , we have to evaluate feynman diagrams directly . by following the results @xcite for standard spurion @xmath157 in @xcite the authors have derived the soft scalar mass spectrum in the model defined in eq.([s2 ] ) , where it is crucial to establish the connection between the mass eigenstates of messenger fields in these two cases . following this insight , we derive the correction due to yukawa superpotential deformation eq.([def ] ) in terms of results in the standard gm with yukawa superpotential deformation eq.([def ] ) . in terms of eq.([s2 ] ) one obtains the mass matrixes @xmath159 and @xmath160 for lepton messenger scalars , and @xmath161 for lepton messenger fermions . the scalar mass matrix in eq.([a3 ] ) is the same as minimal gm , while the scalar mass matrix in eq.([a4 ] ) is diagonal . so only the fermion mass matrix @xmath162 in eq.([a5 ] ) needs to be diagonalized . similar to @xcite we assume an unitary matrix @xmath163 to achieve this , with @xmath164 an angle . in principle , we need two independent unitary matrixes for more general @xmath162 other than that in eq.([a5 ] ) . the two eigenstate masses for fermion messengers are equal to @xmath1 . consequently , we build the connection between gauge and mass eigenstates as , @xmath165 and , @xmath166 the correction to soft scalar mass spectrum due to the yukawa superpotential deformation to the standard gm has been previously studied in @xcite . we find that the the messenger sector is divided into two independent subsectors @xmath167 and @xmath168 , and masses for messenger scalars and fermions in the subsector @xmath167 are degenerate at @xmath1 . so the correction to soft mass spectrum only attributes to subsector @xmath168 , which reads as , @xmath169&\simeq&\mathcal{a } ' \cdot \cos^{2}\theta \cdot \left(-\frac{1}{2}d_{h}\alpha_{t}[m]\alpha_{\lambda}[m]\right),\nonumber\\ \delta m^{2}_{\tilde{u}_{3}}[m]&\simeq&\mathcal{a } ' \cdot \cos^{2}\theta \cdot \left(-d_{h}\alpha_{t}[m]\alpha_{\lambda}[m]\right),\\ \delta m^{2}_{h_{u}}[m]&\simeq&\mathcal{a } ' \cdot \cos^{2}\theta \cdot \left(\frac{1}{2 } d_{h}(d_{h}+3)\alpha^{2}_{\lambda}[m]-d_{h}\alpha_{\lambda}[m]\sum_{r=1}^{r=3}c^{r}\alpha_{r}[m]\right),\nonumber\\ \delta a^{2}_{t}[m]&\simeq&\mathcal{a } ' \cdot \cos^{2}\theta \cdot \left(\frac{1}{2}d^{2}_{h}\alpha^{2}_{\lambda}[m]\right),\nonumber\end{aligned}\ ] ] with @xmath118 , here the structure constants are defined as , @xmath170\equiv\frac{g^{2}_{r}[m]}{4\pi},~~~~~~~~ \alpha_{t}[m]\equiv\frac{y_{t}^{2}[m]}{4\pi},~~~~~~~~ \alpha_{\lambda_{s}}[m]\equiv\frac{\lambda_{s}^{2}[m]}{4\pi},\end{aligned}\ ] ] with @xmath171 ( @xmath172 ) refer to the sm gauge couplings . for the model under study we have @xmath173 and @xmath174 , @xmath108 being number of messenger pairs . @xmath175 is the sum of quadratic casimirs of the fields which participate in the higgs - messenger - messenger yukawa coupling , with @xmath176 referring to messenger fields involved . in our case , @xmath177 . 99 @xmath178 collaboration , `` combined search for the standard model higgs boson using up to 4.9 fb@xmath179 of @xmath180 collision data at @xmath181 tev with the atlas detector at the lhc , '' phys . b710 , 49 ( 2012 ) , arxiv:1202.1408 [ hep - ex ] ; + @xmath182 collaboration , `` combined results of searches for the standard model higgs boson in @xmath180 collisions at @xmath181 tev , '' phys . b710 , 26 ( 2012 ) , arxiv:1202.1488 [ hep - ex ] . j. l. feng , p. kant , s. profumo and d. sanford , `` three - loop corrections to the higgs boson mass and implications for supersymmetry at the lhc , '' phys . * 111 * , 131802 ( 2013 ) , arxiv:1306.2318 [ hep - ph ] . y. nomura and k. tobe , `` phenomenological aspects of a direct transmission model of dynamical supersymmetry breaking with the gravitino mass m(3/2 ) < 1-kev , '' phys . d * 58 * , 055002 ( 1998 ) , [ hep - ph/9708377 ] .	in this paper , we explore conditions for focus point in the high - scale supersymmetry with the weak - scale gaugino masses . in this context the tension between the naturalness and lhc 2013 data about supersymmetry as well as the cold dark matter candidate are addressed simultaneously . it is shown that the observed higgs mass can be satisfied in a wide classes of new models , which are realized by employing the non - minimal gauge mediation .
for univariate polynomial equations defined over a field , it is desirable to obtain general upper bounds on the number of solutions given in simple terms of plainly available information , such as the coefficients , exponents , or number of terms . the ubiquitous example of this is the degree bound , but over non - algebraically closed fields , it is possible to considerably improve upon the degree bound for certain non - negligible families of polynomials . over the real numbers , descartes rule of signs implies that a @xmath8-nomial @xmath9 must have less than @xmath10 real roots . for sparse polynomials - those with a small number of nonzero terms - this can provide a remarkable improvement on the trivial upper estimate given by the degree of @xmath9 . in @xcite , the authors establish a finite field analogue of descartes rule : a sparsity - dependent upper bound on the number of roots of a @xmath8-nomial over @xmath2 . more recently , an improved upper bound was derived in @xcite . here , we investigate possible further improvements to the bound for the special case of @xmath11 . this can be considered the smallest nontrivial choice of @xmath8 , since the zero sets of univariate binomials are easily characterized - they are simply cosets of subgroups of @xmath12 , possibly together with @xmath13 . [ sparse ] ( * ? ? ? * theorems 2.2 and 2.3 ) let @xmath14\ ] ] with all @xmath15 nonzero and @xmath16 . if @xmath9 vanishes on an entire coset of a subgroup @xmath17 , then @xmath18 furthermore , let @xmath19 denote the number of distinct roots of @xmath9 in @xmath20 , and suppose @xmath21 . if @xmath22 denotes the maximal cardinality of a coset on which @xmath9 vanishes , then @xmath23 for a trinomial @xmath24 $ ] , with @xmath25 and @xmath26 nonzero , associate the parameter @xmath27 suppose that @xmath21 . it follows from theorem [ sparse ] that if @xmath9 vanishes on a coset of size @xmath22 , then @xmath28 . since @xmath22 must divide @xmath29 , we have that @xmath22 divides @xmath30 . on the other hand , if @xmath9 vanishes at @xmath31 , then @xmath32 , and @xmath9 vanishes on the entire coset @xmath33 of order @xmath30 . so , in the trinomial case we have explicitly that @xmath34 , and the bound given above simplifies to @xmath35 as pointed out in @xcite , this bound for trinomials is also a consequence of an earlier result from @xcite which bounds the number of cosets @xmath36 needed to express the zero set of a sparse polynomial as a union of the form @xmath37 . our first result refines this upper bound . [ upper_bound ] the roots of a trinomial @xmath38\ ] ] are the union of no more than @xmath39 cosets of the subgroup @xmath17 of size @xmath30 . consequently , we now have @xmath40 , improving the previous result by approximately a factor of @xmath41 when @xmath42 . the method of proof is elementary but interesting : given a trinomial with @xmath43 and @xmath44 roots in a field of undetermined size , we construct @xmath45 distinct nonzero elements in the field , giving a lower bound on its size . additionally , we show that when @xmath6 , this new bound is optimal for even - degree extensions of @xmath46 . if @xmath5 is an even power of a prime @xmath47 and @xmath6 , the bound reduces to @xmath48 , and we can indeed construct trinomials with @xmath6 and @xmath4 distinct roots in @xmath20 . [ june_trinomials ] for any odd prime @xmath47 , the trinomial @xmath49 has exactly @xmath50 roots in @xmath51 . we prove theorem [ june_trinomials ] via linear - algebraic techniques : the extremal examples provided are translations of linear maps with null - spaces of exactly half the dimension of @xmath52 as a vector space over @xmath53 . the optimality of the bound is somewhat murkier when @xmath20 is not an even - degree extension . trinomials with nearly as many roots have been found for some other cases ; for example , when @xmath5 is a cube , the authors of @xcite give the example @xmath54 which has @xmath55 roots . most notably , the question of optimality of the bound remains open for the prime field case . we remark that out of all examples of which we are aware , including those given in @xcite , the only sparse polynomials which vanish at a substantial number of points do so by abusing some obvious algebraic structure of @xmath20 - they either vanish on an entire translation of a subspace or on an entire coset of a nontrivial subgroup . trinomials over prime fields which have @xmath43 are deprived of both of these luxuries , and accordingly , finding examples with many roots seems to be difficult . let @xmath56 denote the maximum value of @xmath19 over all trinomials in @xmath57 $ ] having @xmath43 . recall that by fermat s little theorem , if @xmath58 and @xmath59 , then the two polynomials @xmath60 and @xmath61 define the same mapping on @xmath62 , so it is possible to compute @xmath56 in straightforward way by enumerating trinomials with degree less than @xmath63 and counting their roots . in @xcite , @xmath56 is computed for all primes up to 16633 , and they find no instances in which @xmath56 exceeds @xmath64 . as a result of a large - scale computation , we observe that the inequality @xmath65 continues to hold for all primes up to 139571 . therefore , it appears that the current bound , @xmath66 , is still far from optimal for trinomials over @xmath46 , but we have been unsuccessful in proving any stronger version of theorem [ upper_bound ] for prime fields . it is known that if @xmath9 is allowed to range over all polynomials in @xmath57 $ ] , the distribution of @xmath19 approaches a poisson distribution with mean @xmath67 as @xmath68 @xcite . that is , the proportion of @xmath69 $ ] with @xmath70 is approximately @xmath71 when @xmath47 is sufficiently large . based on computational experiments , this also appears to be true when @xmath9 ranges over just the set of trinomials in @xmath57 $ ] with @xmath6 . this is certainly _ not _ the case if @xmath9 were to range over , for example , the set of _ all _ trinomials , or the set of _ tetranomials _ with @xmath6 due to the presence of @xmath9 which vanish on large cosets . on the other hand , the set of trinomials with @xmath6 appears to behave similarly to what we would expect from an @xmath9 randomly selected from all of @xmath57 $ ] . apparently , restriction of @xmath69 $ ] to trinomial with @xmath43 provides very little statistical information about @xmath19 . [ behave ] with respect to root number , the set of trinomials @xmath69 $ ] with @xmath6 behaves like a uniform random sample of polynomials from @xmath57 $ ] . that is , when @xmath47 is large enough , the values of @xmath19 behave like they are given by random variables with distribution function @xmath72 ( a poisson distribution with mean 1 ) . in section 4 , we show that heuristic [ behave ] allows us to make fairly accurate guesses of the actual values of @xmath56 recorded by our computations . therefore , it may be that the observed logarithmic growth of @xmath56 is not due to any special property of trinomials with @xmath43 , but rather emerges as a statistical consequence of this set being so ordinary , " together with the exponential decay of the poisson distribution . we phrase this formally as the following conjecture that the distributions of @xmath73 and @xmath19 , with @xmath74 ranging over @xmath57 $ ] and @xmath9 ranging over trinomials with @xmath43 , differ by at most a constant factor . [ con ] define @xmath75 \ : : \ : \deg f < p\ } } , \\ t_p & = { \{x^n + ax^s + b \ : : \ : ( a , b ) \in ( { \mathbb{f}}_p^)^2 , \ ; 0 < s < n < p - 1 , \ ; \textnormal{and } \gcd(n , s , p-1 ) = 1\}}.\end{aligned}\ ] ] let @xmath76 denote the proportion of @xmath77 with @xmath70 and @xmath78 denote the proportion of @xmath79 with @xmath70 . there exists a constant @xmath80 such that @xmath81 for all @xmath47 prime and @xmath82 . if these two distributions do in fact differ by at most a constant factor , then we are able to readily derive the logarithmic upper bound for @xmath56 suggested by our experiments . and , in turn , we could extend such a bound to trinomials with @xmath83 by noticing that @xmath84 has at most @xmath30 roots for every root of @xmath85 . [ cor ] suppose conjecture [ con ] is true . then , we have the asymptotic bound @xmath86 let @xmath87 and @xmath88 so that @xmath89 we can bound @xmath90 from above by counting polynomials of the form @xmath91 with @xmath92 distinct , which gives @xmath93 obviously @xmath94 , so assuming the existence of @xmath95 defined in conjecture [ con ] , we have @xmath96 if @xmath97 then the set @xmath98 is empty , so we must have @xmath99 , or equivalently , @xmath100 by applying stirling s approximation , we get the asymptotic bound @xmath101 by considering the growth order of the inverse function of @xmath102 , we obtain @xmath103 we remark that in @xcite , it is shown that under the generalized riemann hypothesis , there exists an infinite sequence of primes @xmath104 satisfying the lower bound @xmath105 . so , the truth of both conjecture [ con ] and grh would imply that the bound in corollary [ cor ] is , up to a multiplicative constant , asymptotically optimal . finally , we prove the following theoretical result which states that conjecture [ con ] is true if we consider only trinomials of bounded degree as we take @xmath47 to infinity . this weaker result is suggestive but certainly not sufficient to imply the bound in corollary [ cor ] . in particular , theorem [ poisson_density ] shows the existence of @xmath106 such that @xmath107 , but we do not have a bound on the set @xmath108 . [ poisson_density ] suppose @xmath109 with @xmath110 and @xmath111 . as @xmath68 , the proportion of pairs @xmath112 , such that @xmath60 has @xmath70 , converges to @xmath113 } { \hfill r ! ( n - r ) ! \ } & \mbox{if } r < n \\ \ : 1/r ! & \mbox{if } r = n , \end{dcases}\ ] ] where @xmath114 $ ] denotes the nearest integer " function . + furthermore , fix @xmath115 , and let @xmath116 let @xmath117 be defined as in conjecture [ con ] , and let @xmath118 denote the proportion of @xmath119 with @xmath70 . we then have @xmath120 for @xmath121 fixed , define the family of trinomials in @xmath122 $ ] @xmath123 observe that @xmath124 is exactly the set of trinomials with support @xmath125 * constant term @xmath67 * @xmath126 this is clear because @xmath126 if and only if @xmath9 s coefficients sum to zero . we introduce this family of trinomials because they have the following useful property . [ disjoint_roots ] let @xmath127 be the unique multiplicative subgroup of order @xmath128 , and suppose that @xmath129 . the only root in @xmath130 shared by any two members of @xmath124 is @xmath131 . @xmath132 is equivalent to the following linear equation in @xmath133 : @xmath134 this has multiple solutions in @xmath133 if and only if both @xmath135 and @xmath136 . since @xmath130 is a cyclic group of order @xmath128 and @xmath129 , the only @xmath137 such that @xmath138 is @xmath131 itself . so 1 is the only @xmath139 such that @xmath132 for multiple @xmath140 . [ gcd_one ] let @xmath127 be the unique multiplicative subgroup of order @xmath128 , and let @xmath141 $ ] be a trinomial of the form @xmath142 satisfying @xmath129 . the number of roots of @xmath9 that lie in @xmath130 does not exceed @xmath143 suppose @xmath144 has @xmath44 distinct roots @xmath145 in @xmath130 . for each @xmath146 let @xmath147 . since the map @xmath148 permutes the elements of @xmath130 , each of these @xmath149 also has @xmath44 roots in @xmath130 . additionally , each @xmath149 is a member of @xmath150 , since @xmath151 . we now check that each @xmath149 is distinct . suppose @xmath152 with @xmath153 . we then have both @xmath154 and @xmath155 , or , equivalently , @xmath156 and @xmath157 . once again , the only @xmath137 that satisfies @xmath158 is @xmath131 , so @xmath159 , which contradicts the supposition that the roots @xmath145 are distinct . in summary , there exist @xmath44 distinct trinomials of @xmath150 that each have @xmath44 roots in @xmath130 , and by lemma [ disjoint_roots ] , @xmath160 is the only root among these that is not unique . this implies that @xmath130 contains at least @xmath161 distinct elements , but we know that @xmath130 has size @xmath128 by hypothesis . therefore it must be that @xmath162 which yields the desired constraint on @xmath44 : @xmath163 we now have everything we need to complete the proof . let @xmath24 $ ] . obviously the roots of @xmath9 are not affected by re - scaling ; let @xmath164 the exponents may fail to satisfy @xmath165 . however , @xmath166 is equivalent to the system @xmath167 @xmath168 the second equation is only solvable for @xmath169 when @xmath170 lies in the subgroup of order @xmath171 . the first equation satisfies @xmath172 , so we can invoke lemma [ gcd_one ] and find that there are at most @xmath173 such @xmath170 . each of these @xmath170 then admits one coset of @xmath30 distinct solutions for @xmath169 . observe that the function @xmath174 is an @xmath175-linear map from @xmath176 to @xmath176 . @xmath177 is a binomial , so it is easy to show that it does have nonzero solutions and therefore has a null space of positive dimension . since @xmath177 is not the zero transformation , we conclude that it has a null space of dimension 1 , and therefore that it has @xmath50 roots . we see that @xmath178 exactly when @xmath179 . this has one obvious solution , @xmath180 , so we conclude from the linearity of @xmath177 that it has as many solutions as @xmath181 . therefore , @xmath9 has @xmath182 roots , all of which are nonzero . our proof of theorem [ poisson_density ] relies on the following statement from @xcite , which can be viewed as a chebotarev density theorem for function fields . at its core , this result is powered by the lang - weil estimate for the number of points on varieties over @xmath20 . [ ffc ] ( * ? ? ? * proposition 3.1 ) let @xmath183 and @xmath128 be positive integers , and let @xmath184 $ ] be separable in @xmath169 and have @xmath185 and @xmath186 . let @xmath187 be an algebraic closure of @xmath20 , and suppose that @xmath188 for a partition @xmath95 of @xmath189 , let @xmath190 denote the conjugacy class of permutations @xmath191 with cycle type @xmath95 , and let @xmath192 denote the set of @xmath193 such that the univariate polynomial @xmath194 factorizes over @xmath20 into irreducible factors with degree pattern @xmath95 . then , there exists a constant @xmath195 , which depends only on @xmath196 and @xmath128 , such that @xmath197 let @xmath198 be a field , and let @xmath199 , where @xmath200 and @xmath201 are indeterminates over @xmath198 , @xmath110 , and @xmath111 . it is shown by cohen in ( * ? ? ? * and corollary 3 ) that unless char@xmath202 divides @xmath203 , @xmath74 is separable over @xmath204 and @xmath205 . here we consider @xmath198 an algebraic closure of a prime field @xmath46 , and @xmath189 bounded by some fixed @xmath115 , so @xmath74 satisfies the conditions of theorem [ ffc ] when @xmath206 . let @xmath207 be the collection of all permutations @xmath191 with exactly @xmath44 fixed points . if @xmath208 then @xmath207 contains only the identity permutation and then @xmath209 . otherwise , every @xmath210 can be written as @xmath211 , where each @xmath15 is a length - one cycle and @xmath212 permutes the remaining elements and has no fixed points . permutations that have no fixed points are called _ derangements _ , and the proportion of permutations that are derangements is extremely well - approximated by @xmath213 @xcite . specifically , the number of derangements of @xmath189 elements is given by @xmath214,$ ] where @xmath114 $ ] denotes the nearest integer " function . therefore , to count the the number of @xmath210 , we simply count the ways to choose @xmath215 and multiply by the number of derangements of the remaining @xmath216 elements , so we have @xmath217}{r!(n - r)!}.\ ] ] note that @xmath218}{r!(n - r ) ! } \leq \frac { e^{-1}(n - r ) ! + 0.5}{r!(n - r ) ! } \leq \frac { e^{-1}+ 0.5}{r ! } < \frac{1}{r!},\ ] ] so in fact we have @xmath219 always . let @xmath220 denote the number of @xmath221 such that @xmath222 $ ] has exactly @xmath44 linear factors . since @xmath207 is the union of some number of conjugacy classes which is bounded in terms of @xmath128 , we have @xmath223 as @xmath68 , where the @xmath224-constant depends only on @xmath128 . now define @xmath225 @xmath226 differs negligibly from @xmath220 since there are less than @xmath227 elements in @xmath228 , and by ( * ? ? ? * proof of proposition 3.1 ) , the number of @xmath221 such that @xmath84 has a root of multiplicity is bounded asymptotically by @xmath229 , so @xmath230 finally , note that @xmath231 therefore we have @xmath232 which proves the first claim , concerning trinomials with @xmath111 . + recall the definitions @xmath75 \ : : \ : \deg f < p\ } } \\ t_{p , n } & = { \{x^n + ax^s + b \ : : \ : ( a , b ) \in ( { \mathbb{f}}_p^)^2 , \ ; 0 < s < n \leq n , \ ; \textnormal{and } \gcd(n , s , p-1 ) = 1\}},\end{aligned}\ ] ] and recall that @xmath76 denotes the proportion of @xmath77 with @xmath70 , and that @xmath118 denotes the proportion of @xmath119 with @xmath70 . it is clear that @xmath118 is equal to the average value across all fractions @xmath233 which are associated to a trinomial @xmath234 with @xmath235 and @xmath236 . it remains to study trinomials with @xmath237 but @xmath238 . suppose @xmath239 , and write @xmath240 and @xmath241 so that @xmath242 . if @xmath237 , then we must have @xmath243 , so the map @xmath244 permutes @xmath46 . since @xmath245 and @xmath246 , it follows that the trinomials @xmath84 and @xmath247 have the same number of distinct roots in @xmath46 . thus , @xmath118 is equal to the average of a collection of fractions which all satisfy @xmath248 where @xmath249 is a constant which depends only on @xmath128 . it follows immediately that @xmath250 for each @xmath251 , and so @xmath252 in @xcite , leontev studies the generating function @xmath253 , and shows that @xmath254 converges to @xmath255 for @xmath256 $ ] . using the continuity theorem for generating functions ( ? ? ? * section 1.1.6 ) , he then concludes that @xmath257 as @xmath68 for all @xmath82 . since we are only interested in the finitely many @xmath258 , we can also be assured that @xmath259 therefore , we have @xmath260 first , we attempt to establish some basic plausibility for the poisson heuristic . as before , let @xmath78 denote the proportion of trinomials over @xmath46 with @xmath43 that have @xmath44 distinct roots . the following table gives the statistical distance between @xmath8 and a poisson distribution with mean @xmath67 for a few fields of various sizes . + [ cols="<,^",options="header " , ] recall that @xmath261 denotes the set of trinomials over @xmath46 with @xmath43 and degree less than @xmath63 . we have computed @xmath56 , the maximum number of roots of attained by any @xmath79 , for primes up to @xmath262 . in this section , we show that the values of @xmath56 that we would expect by heuristic [ behave ] are quite close to what we actually observe . that is , we consider the expected values of @xmath263 under the model that the values of @xmath19 are given by random variables with distribution function @xmath264 , and we compare these expected values with real values of @xmath56 . more generally , let @xmath265 be the maximum of @xmath128 independent variables all with distribution @xmath264 . it is known that @xmath265 becomes very predictable when @xmath128 is large . specifically , it is shown in @xcite that there exists an integer sequence @xmath266 such that , as @xmath267 , @xmath268 in @xcite , a nice asymptotic formula is given for @xmath266 : @xmath269 as an initial estimate , there are slightly less than @xmath270 trinomials @xmath271 : there are @xmath272 pairs @xmath273 and almost @xmath272 pairs @xmath274 . so , assuming heuristic [ behave ] , a reasonable conservative prediction would be @xmath275 however , to make an accurate prediction for @xmath56 we need to be more precise in two ways . firstly , there are actually much fewer independent values of @xmath19 than @xmath270 . for any @xmath69 $ ] , we have that @xmath276 as long as @xmath277 and @xmath278 , because the maps @xmath279 and @xmath280 are both bijections on @xmath46 . as a result , knowing the number of roots of one trinomial immediately determines the number of roots of a significant chunk of trinomials . therefore , we would like to find an appropriate , effective value for @xmath128 that better models the number of independent random values . to do this , we count exactly the number of trinomials with @xmath43 and then quotient out by the size of these equivalent chunks . the exact number of pairs @xmath281 that are relatively prime with @xmath63 is given by the _ jordan totient function _ , @xmath282 @xcite . we must subtract @xmath283 to avoid counting pairs with @xmath284 , and we divide by @xmath41 to avoid counting both @xmath274 and @xmath285 . there are @xmath272 choices for the two coefficients , so overall we have @xmath286 as discussed in section 2 , @xmath287 implies that every pair @xmath288 is unique , so we divide by @xmath289 to account for trinomials of the form @xmath290 . to account for the transformation @xmath280 , we divide by the number of @xmath291 with @xmath292 , which is given by @xmath293 . so , we take our effective number of independent poisson variables to be @xmath294 this number is approximately equal to @xmath295 ; for primes in the range @xmath296 , we have @xmath297 secondly , it is beneficial to consider the less elegant but more precise asymptotic formula for @xmath266 given in @xcite . below , @xmath298 is the _ lambert w function_. @xmath299 in summary , by heuristic [ behave ] we expect that @xmath300 when @xmath47 is sufficiently large . the following plot displays the ratios @xmath301 for all primes @xmath302 . the visibly distinct bands correspond to primes that share the same value for @xmath56 . the apparent upper and lower bounding monotonic subsequences are traced by dotted curves . the average over all ratios is 1.0429 and the standard deviation is 0.05587 . for all @xmath302 , we have @xmath65 . the largest recorded value of @xmath56 is @xmath303 , which is witnessed at @xmath304 and @xmath305 ; the associated ratios @xmath306 lie visibly on the upper dotted line . the values of @xmath307 were computed in a straightforward way ( i.e. by enumerating trinomials and counting their roots ) by parallel c++ code which ran on texas a&m s ada supercomputing cluster for @xmath308 cpu hours . the program takes advantage of the fact that @xmath309 when @xmath292 and @xmath278 to reduce the enumeration space . the values @xmath310 were computed separately by a small matlab program , which in particular makes use of matlab s built - in ` lambertw ` function . we would like to thank the texas a&m supercomputing facility for providing us with computational resources , and our advisor , j. maurice rojas , for his indispensable guidance . anderson , c.w . , 1970 . extreme value theory for a class of discrete distributions with applications to some stochastic processes . _ journal of applied probability _ , bank , e. , bary - soroker , l. and rosenzweig , l. , 2015 . prime polynomials in short intervals and in arithmetic progressions . _ duke mathematical journal , 164_(2 ) , pp.277 - 295 . bi , j. , cheng , q. and rojas , j.m . , 2013 , june . sub - linear root detection , and new hardness results , for sparse polynomials over finite fields . _ proceedings of the 38th international symposium on symbolic and algebraic computation _ 61 - 68 ) . briggs , k.m . , song , l. and prellberg , t. , 2009 . a note on the distribution of the maximum of a set of poisson random variables . _ arxiv preprint arxiv:0903.4373_. canetti , r. , friedlander , j. , konyagin , s. , larsen , m. , lieman , d. and shparlinski , i. , 2000 . on the statistical properties of diffie - hellman distributions . _ israel journal of mathematics , 120 _ , pp.23 - 46 . cheng , q. , gao , s. , rojas , j.m . and wan , d. , 2014 . sparse univariate polynomials with many roots over finite fields . _ arxiv preprint arxiv:1411.6346_. cohen , s. , 1980 . the galois group of a polynomial with two indeterminate coefficients . _ pacific journal of mathematics , 90_(1 ) , pp.63 - 76 . dickson , l.e . , 1919 . _ history of the theory of numbers , vol . i _ , carnegie institution , washington dc . kelley , z. , 2016 . roots of sparse polynomials over a finite field . _ arxiv preprint arxiv:1602.00208_. kimber , a.c . , 1983 . a note on poisson maxima . _ probability theory and related fields , 63_(4 ) , pp.551 - 552 . hassani , m. , 2003 . derangements and applications . _ journal of integer sequences , 6_(2 ) , p.3 . leontev , v.k . , 2006 . roots of random polynomials over a finite field . _ mathematical notes , 80_(1 ) , pp.300 - 304 . sachkov , v.n . , 1997 . _ probabilistic methods in combinatorial analysis_. cambridge university press , cambridge .	we show that univariate trinomials @xmath0 $ ] can have at most @xmath1 distinct roots in @xmath2 , where @xmath3 . we also derive explicit trinomials having @xmath4 roots in @xmath2 when @xmath5 is square and @xmath6 , thus showing that our bound is tight for an infinite family of finite fields and trinomials . furthermore , we present the results of a large - scale computation which suggest that an @xmath7 upper bound may be possible for the special case where @xmath5 is prime . finally , we give a conjecture ( along with some accompanying computational and theoretical support ) that , if true , would imply such a bound .
density functional theory has become a popular computational method in quantum chemistry , because of its ability to handle large molecules accurately but relatively inexpensively@xcite . this success is based on the availability of reliable accurate approximate functionals , and there is a constant need for still further accuracy . the goal of atomization energy errors being reliably less than 1 kcal / mol has not yet been achieved . an important step forward in this search for accuracy came when becke mixed a fraction of exact exchange with a generalized gradient approximation ( gga ) , and reduced errors by a factor of two or three@xcite . such hybrid functionals , e.g. , b3lyp , are now in common use , but their underlying justification comes from the adiabatic decomposition of density functional theory . initially , the mixing parameters used were determined empirically . later , it was shown that these parameters could be derived non - empirically@xcite , and that a single universal mixing coefficient ( 25% ) could be rationalized based on the performance of mp theory for molecules@xcite . most recently , new functionals have been proposed which use this adiabatic decomposition in much detail@xcite . the adiabatic decomposition of an electronic system is very simple conceptually . imagine multiplying the electron - electron repulsion by a coupling constant @xmath0 . now imagine varying @xmath0 , while keeping the electron density @xmath1 fixed . this differs from traditional perturbative methods , e.g. , mller - plesset @xcite , because the external potential must be altered at each @xmath0 to keep the density fixed . at @xmath2 , we have the physical , interacting electronic system . but as @xmath0 is reduced to zero , keeping the density fixed , the electrons become those of the non - interacting , kohn - sham system , and the potential morphs into the kohn - sham potential . all kohn - sham dft calculations are actually performed on this non - interacting system , and the physical ground - state energy deduced from it . the adiabatic connection provides a continuous connection between the interacting system and its kohn - sham analog . the only part of the total energy which must be approximated in such a kohn - sham calculation is the exchange - correlation energy , @xmath3 $ ] , as a functional of the ( spin ) density . a further value of the coupling constant is that , through the hellmann - feynman theorem , this energy can be written as an intergal over the purely potential contribution@xcite : e= _ 0 ^ 1 d u ( ) , u()=\|v\| - u , [ acm ] where @xmath4 is the coulomb interaction between electrons , @xmath5 is the wavefunction at coupling constant @xmath0 , and @xmath6 is the hartree electrostatic energy . this integral is the adiabatic connection formula . hybrid functionals are based on the fact that this integrand , when applied to the exchange - correlation contribution to a bond dissociation energy , is often _ not _ well - approximated by gga s , due to their lack of static correlation@xcite . this can be partially corrected for most molecules by mixing in a fraction of exact exchange . construction of functionals based on this insight is often referred to as the adiabatic connection method ( acm ) . ( 12,6.5 ) ( -5.2,4.0 ) ( 12,6.5 ) 6= ( 0.8,3.85 ) ( 0,0)6 ( 5.1,0.1)@xmath0 ( 2.8,3.0)exact ( 4.5,4.0)pbe [ uxcl ] thus accurate approximation of this adiabatic connection curve is extremely important to further progress in construction of approximate functionals , and benchmark cases for small systems are always of interest and help in this endeavor . however , this is , in principle , a very demanding task . for each value of @xmath0 , one must solve the interacting electronic problem many times in order to find the external potential which reproduces the @xmath2 density . almbladh and pedroza@xcite made early heroic attempts , but not with the accuracy of modern calculations . in the last several years , as the importance of these curves has become apparent , several groups in different areas have performed adiabatic decomposition calculations . hood et al.@xcite have calculated the curve for bulk si . colonna and savin@xcite have used the slightly different harris and jones@xcite decomposition for both the he and be isoelectronic series . joubert and srivastava@xcite have used hylleras - type wavefunctions to calculate these curves for the he isoelectronic series . all these methods require solving the interacting electronic problem with a potential that differs from the original external potential , e.g. , @xmath7 for atoms . this typically makes them difficult to transfer to other systems , radically different from the original one , in which other codes and approximations are being used . in this paper , we show how to construct the adiabatic connection formula for any atom accurately by doing calculations simply for different values of @xmath8 , so that no modification need be made to an existing wavefunction code . a standard calculation is simply run for several different nuclear charges , running from the physical value up to @xmath9 . the main focus of this paper is to demonstrate the accuracy of the method . the solid line in figure 1 is an essentially exact curve for @xmath10 for hooke s atom ( two electrons in a harmonic potential ) , for a spring constant of @xmath11 ( atomic units ) , calculated from a series of exact calculations for force constants greater than 1/4 , but never requiring any calculation with a different external potential . the method can be immediately applied to accurate calculations for larger atoms , and is currently being explored for molecules . the paper is divided into several sections . in the next section , we outline the basic theory behind our calculations . following that , we present results , both using the exact functional , and within a gga , for both hooke s atom and the he atom . hooke s atom consists of two electrons attached by springs to a center , but interacting via a coulomb repulsion . we summarize our findings in the last section . atomic units ( @xmath12 ) are used throughout , and only spin - unpolarized systems are discussed . for convenience , we begin with a review of the adiabatic connection formalism . we define @xmath13 $ ] to be that wavefunction which has density @xmath1 and minimizes f [ ] = \| t + \| . [ fl ] then the kinetic - correlation energy at @xmath0 is t= \| t \| - t[tlc ] where @xmath14 $ ] is the non - interacting kohn - sham kinetic energy of density @xmath15 . the potential contribution to exchange - correlation at coupling constant @xmath0 is , u= \|v\| - u [ ] , [ uxcldef ] while the exchange - correlation energy at @xmath0 is e= u+ t . [ excl ] for @xmath2 , the system becomes the true interacting quantum - mechanical system , and energies without superscripts refer to @xmath2 , e.g. , @xmath16 . all quantities at coupling constants different from one are simply related to their full coupling strength counterparts , but evaluated on a scaled density@xcite . the most important examples are the wavefunction : [ ] = _ , [ psil ] and the exchange - correlation energy e [ ] = ^2 e , [ excscal ] [ fund ] where @xmath17 , and @xmath18 @xmath19 . thus knowledge of how a quantity varies as the density is scaled implies knowledge of its coupling constant dependence . quantities evaluated on the kohn - sham wavefunction vary in a simple way , such as the non - interacting kinetic energy @xmath20 : t [ ] = t , or t= ^2 t , [ tsscal ] the hartree electrostatic energy , u[]=u [ ] , or u[]= u [ ] , [ uscal ] and the exchange energy e [ ] = e , or e = e . [ exscal ] correlation is more sophisticated . note , however , that knowledge of any quantity , @xmath21 , t\c [ \n\g]$ ] , or @xmath22 $ ] as a function of @xmath23 for @xmath23 between 1 and @xmath24 , i.e. , scaling to the high density limit , is sufficient to determine the adiabatic connection for any of them , for @xmath0 between 0 and 1 . the most well - known relation is to extract the kinetic - correlation piece from @xmath25 $ ] @xcite : tc [ ] = - ec [ ] + . [ tcfromec ] to generalize this result to @xmath26 , apply eq . ( [ tcfromec ] ) to @xmath26 , and make a change of variables in the derivative , to find : tc [ ] = - ec [ ] + . [ tcfromecg ] this can be considered a first - order differential equation in @xmath23 for @xmath27 $ ] . solution of this equation yields ec [ ] = - _ ^ tc[_ ] , [ ecfromtcg ] where we have used the fact that @xmath27 $ ] has vanished as @xmath28 @xcite . via eq . ( [ fund ] ) , these can be turned into coupling - constant relations : t= e- , [ tclfromecl ] and e= - _ 0^ tc^. [ eclfromtcl ] the latter is called bass relation@xcite . similarly , we can extract @xmath29 from @xmath27 $ ] , since @xmath30 . ( [ tcfromecg ] ) leads to uc [ ] = 2 ec [ ] - [ ucfromecg ] which , inverted , is ec[]=^2 _ ^ uc[_ ] . [ ecfromucg ] combining eqs . ( [ ucfromecg ] ) and ( [ ecfromtcg ] ) then leads to simple relations between @xmath29 and @xmath31 . lastly , the coupling - constant relation that follows from eq . ( [ ecfromucg ] ) is ec [ ] = _ 0 ^ 1 u [ ] , [ ecfromucl ] and @xmath32 of fig . 1 is just the integrand when this expression is applied to both exchange and correlation , i.e. , @xmath33 . we emphasize that all these relations follow from the well - known eqs . ( [ fund ] ) and ( [ tcfromec ] ) . we begin this section by noting how , when some parameter in an external potential is altered , the density changes scale , but often does not change shape very much . for example , for the two - electron ion , going from z=2 to z=4 will roughly multiply the density by 8 , and reduce its length scale by a factor of 2 . we use this fact to very accurately approximate scaling the density . then , through the relations derived above , we can convert this into the coupling - constant dependence . let @xmath1 be the density of the system we are interested in . suppose we alter some parameter in the external potential , and solve the interacting problem , finding some density @xmath34 . if we want to treat @xmath35 as an approximation to @xmath26 , we must first choose a criterion for determining @xmath23 . in fact , either of eqs . ( [ tsscal ] ) and ( [ uscal ] ) could be used , since both would be satisfied exactly if @xmath35 were truly a scaled density . for definiteness , we choose = 1/=u[]/u[ ] . [ ldef ] we can then consider @xmath36 \approx e\c[\n\g]$ ] , @xmath37 \approx t\c[\n\g]$ ] , etc . , which we call the bare estimates . as we shall show in the results section , these energies usually yield quite accurate approximations to the exact quantities . however , since we are not working with the true scaled density , these energies do not satisfy the relations involving scaling derivatives above , such as eq . ( [ ecfromtcg ] ) . to make a better estimate , we note that , in the case of the correlation energy , we can calculate the leading correction to @xmath27 $ ] , using the correlation potential . this can be constructed by finding the exact kohn - sham potential which generates the density , and subtracting from it the external , hartree , and exchange contributions . then , ec [ ] ec[ ] + d^3r vc[ ] ( ) ( ( ) - ( ) ) + o()^2 . [ eccorr ] as will be shown below , inclusion of this correction leads to extremely accurate adiabatic connection curves . we generate @xmath38 $ ] using eq . ( [ ecfromtcg ] ) , and then @xmath30 . in particular , as @xmath39 , @xmath40 , and eq . ( [ eccorr ] ) becomes exact to first order in the difference between @xmath35 and @xmath15 . thus eq . ( [ tcfromec ] ) is satisfied exactly . lastly , we apply the same principles to @xmath20 , as a test of the closeness of the approximate density , @xmath35 . the euler equation for the kohn - sham system says that = - v ( ) , [ tspot ] where @xmath41 is a constant , and @xmath42 is the kohn - sham potential . thus , we can also correct the bare @xmath20 estimate , to t= t- d^3r v ( ) ( ( ) - ( ) ) + o()^2 . ( 12,6.5 ) ( -5.2,4.0 ) ( 12,6.5 ) 6= ( 0.8,3.85 ) ( 0,0)6 ( 5.1,0.1)@xmath43 [ dens ] we present results for two systems , hooke s atom and the he atom . the first of these contains two electrons in a harmonic potential , interacting via a coulomb repulsion . this provides a valid test case for density functional theory , because of the coulomb repulsion between electrons . it is also an easy system to perform calculations on , because the center - of - mass and relative coordinates separate , leaving only a one - dimension differential equation to solve numerically@xcite . even this equation has an analytic solution for force constant @xmath11 @xcite , which is the system on which we will demonstrate most of our results . to implement our method for the @xmath11 hooke s atom , we run many calculations at different values of @xmath44 , up to @xmath45 , using the method of ref . for each calculation , we scale the density so that it looks like the original @xmath11 density , using eq . ( [ ldef ] ) to define the scale factor . several such densities are shown in fig . 2 , which illustrates how close these densities are . note that the largest error is for @xmath46 , the non - interacting gaussian density . ( 12,6.5 ) ( -5.2,4.0 ) ( 12,6.5 ) 6= ( 0.8,3.85 ) ( 0,0)6 ( 5.1,0.1)@xmath0 ( 5.0,4.0)@xmath47 $ ] ( 5.0,2.1)@xmath48 $ ] ( 5.0,5.8)@xmath49 $ ] [ ecpbe ] figure 3 is a plot of the three quantities @xmath47 $ ] , @xmath49 $ ] , and @xmath48 $ ] , all found using the pbe correlation functional . we choose these quantities to plot , as they remain finite in the range @xmath46 to @xmath50 . for each quantity , there are two curves . the dashed curve is the bare result of the calculations at different values of the force constant , i.e. , using @xmath51 $ ] alone . the solid curve for @xmath52 $ ] includes the correction due to the potential , in accordance with eq . ( [ eccorr ] ) . the @xmath29 and @xmath31 curves are then extracted from this one , using eqs . ( [ tcfromecg ] ) and ( [ ucfromecg ] ) . both the value and first derivative at @xmath2 are _ exact _ for this curve . what is more remarkable is that these curves coincide essentially exactly with the correct result , which we can deduce by directly applying eq . ( [ fund ] ) to the pbe correlation energy functional . careful numerical calculations indicate that the maximum error in our curve is at @xmath46 , and is less than 3@xmath53 hartrees . this is due to the similarity of densities , as shown in fig . 2 , leading to very small corrections . we also tested other possible prescriptions for choosing @xmath0 , such as from the square root of the ratio of non - interating kinetic energy densities , as in eq . ( [ tsscal ] ) . this gives values for @xmath0 very close to those of eq . ( [ uscal ] ) , and leads to no measurable change in our estimate for @xmath47 $ ] . ( 12,6.5 ) ( -5.2,4.0 ) ( 12,6.5 ) 6= ( 0.8,3.85 ) ( 0,0)6 ( 5.1,0.1)@xmath0 ( 5.0,4.1)@xmath47 $ ] ( 5.0,2.7)@xmath48 $ ] ( 5.0,5.8)@xmath49 $ ] [ ecexa ] figure 4 is the analog of fig . 3 , but now calculated using exact correlation energies and potentials . based on the remarkable accuracy of fig . 3 , we claim these curves are essentially exact . in fact , the change due to the potential correction is much smaller for the exact case than for pbe , suggesting that the error made in eq . ( [ eccorr ] ) will also be smaller . the adiabatic curve of fig . 1 was derived from these curves , since @xmath54 $ ] . we also ran pbe calculations for lower densities , where the shape of the density can change significantly with @xmath55 . we find that even as low as @xmath56 , there is a maximum error of only 1 millihartree in the corrected @xmath47 $ ] curve . beyond this point , the reliability of our method might be questioned . however , defining an average @xmath57 value via eq . ( 6a ) of ref . @xcite , at this point @xmath58 , whereas for @xmath11 , @xmath59 is 2.07 . thus for common values of the density , the errors in our procedure are minute . ( 12,6.5 ) ( -5.2,4.0 ) ( 12,6.5 ) 6= ( 0.8,3.85 ) ( 0,0)6 ( 5.1,0.1)@xmath0 [ tstest ] as a final test of our method , we return to the non - interacting kinetic energy . if the scaling were exact , the non - interacting kinetic energy would scale quadratically , as in eq . ( [ tsscal ] ) . in figure 5 , we plot @xmath60/\gamma^2 $ ] exactly , approximating @xmath60 $ ] by @xmath61 $ ] , and including the correction of eq . ( [ tscorr ] ) . note that the maximum absolute error , at @xmath46 , is only about -1.6% , in the bare estimate . the correction makes the derivative with respect to @xmath0 exact as @xmath62 , and overall reduces the error by about a factor of 5 , to about -0.3% . ( 12,6.5 ) ( -5.2,4.0 ) ( 12,6.5 ) 6= ( 0.8,3.85 ) ( 0,0)6 ( 5.1,0.1)@xmath0 ( 4.1,2.1)bare estimate ( 2.1,4.9)corrected est . ( 5.1,4.1)exact [ echepbe ] finally , we discuss our results for the two - electron ion series , simply to demonstrate that there is nothing special about hooke s atom which makes these techniques accurate . we may make bare estimates of the adiabatic curves directly from energy data already in the literature . thus by calculating @xmath0 by eq . ( [ ldef ] ) using the data in table i of ref . @xcite , and assigning to each @xmath0 the bare @xmath29 value from the table , we find the corresponding adiabatic connection curve integrates to @xmath63 -41.5 mh , as compared with an exact value of -42.1 mh for he . we can similarly use the @xmath31 data in eq . ( [ ecfromtcg ] ) to get an estimate of -41.2 mh , so that the difference between these two results is a good indicator of the error in both of them . again , we can calculate @xmath64 , and find the largest error at @xmath46 , with a value of 2.916 , as compared to the exact value of 2.867 . in fig . 6 , we plot pbe calculations of @xmath47 $ ] for the he atom , both exactly and using the fake adiabatic connection formula , both with and without the correction term . we use the highly accurate densities of umrigar and gonze@xcite . unfortunately , not enough data points are available to make smooth plots for this system . already at @xmath65 , @xmath66 , so that results for non - integer values of @xmath8 ( between @xmath67 and @xmath68 ) are needed . however , enough data is present to see the clear correction the potential makes . we note several interesting features of this curve . first , the dependence on scaling is much less for he than for the @xmath11 hooke s atom . second , the pbe curve contains a minimum , so eq . ( [ tcfromecg ] ) implies @xmath69 becomes positive for this scaled density . while it has never been rigorously proven , every known case of @xmath69 is negative@xcite . thus this is probably a ( very slight ) limitation of pbe , which does not occur in pw91 , and may be related to the difference at @xmath70 in fig . third , the potential correction even picks up this feature , and still only has errors of a fraction of a millihartree . ( 12,6.5 ) ( -5.2,4.0 ) ( 12,6.5 ) 6= ( 0.8,3.85 ) ( 0,0)6 ( 5.1,0.1)@xmath0 ( 5.1,3.1)corrected est . ( 2.1,4.9)bare estimate [ eche ] figure 7 repeats fig . 6 , but for the exact case . here the corrected curve is much straighter than for pbe , and has no minimum . the lines are drawn merely to aid the eye . note that the value of @xmath71 $ ] , which is the high - density limit of @xmath72 , drops about 1 millihartree , due to the difference in shape between the he atom and a pure exponential decay ( bare estimate ) . we find @xmath73= -47.9 millihartree , in reasonable agreement with the value of colonna and savin@xcite ( -47.5 from table vi ) , of joubert and srivastava@xcite ( -47.2 from @xmath74 in table ii ) , and of engel and dreizler@xcite ( -48.2 from table v ) . the value -50.3 , reported in ref . @xcite , was evaluated on the pbe selfconsistent density . we have shown how , with accurate ground - state results as a function of the external potential , an accurate adiabatic connection curve can be calculated for both hooke s atom at @xmath11 and the he atom . we see no reason why similar results could not be obtained for larger atoms , especially those for which the kohn - sham potential has already been calculated@xcite . there also exist methods for isolating the correlation potential@xcite . in the event that the exchange contribution can not be easily isolated , the scheme can still be applied , but now to the combined exchange - correlation energy at each value of the external parameter . the correction is now evaluated using the exchange - correlation potential . on hooke s atom , this yields results almost as accurate as those described for the correlation energy alone . important differences are in the regime @xmath75 , as here the exchange contribution to the correction blows up , since @xmath76 . this effect causes noticeable errors only for @xmath77 in @xmath47 $ ] , but not in @xmath29 or @xmath31 . it also means that neither @xmath78 $ ] or @xmath79 $ ] , as derived from the resulting @xmath27 $ ] curve via eqs . ( [ ucfromecg ] ) and ( [ tcfromecg ] ) , is exact at @xmath2 . however , these differences are of the order of 0.1 millihartree . finally , we are currently investigating if this method can be used to calculate accurate adiabatic connection curves for molecules@xcite . this work was supported by an award from research corporation and by the national science foundation under grant no . we thank john perdew for his comments , and cyrus umrigar for providing us with accurate densities for the two electron ions .	an extremely easy method for accurately calculating the adiabatic connection of density functional theory is presented , and its accuracy tested on both hooke s atom and the he atom . the method is easy because calculations are needed only for different values of parameters in the external potential , which can be achieved with almost any electronic structure code . psfig ^ # 1 var(_1 ... ) _ c _ h _ int d^3 r rotate # 1 \|_=1 # 1
the schmidt - kennicutt law ( schmidt 1959 ; kennicutt 1998a , b ) , hereafter star formation or sf law , in galaxies has been extensively investigated by measuring surface densities of star formation rate ( sfr ) and surface or column gas densities projected on the galactic planes ( schmidt 1959 ; kennicutt 1998a , b ; komugi et al . 2006 ; bigiel et al . 2008 ; leroy et al . 2008 , 2013 ; lada et al . 2012 ; kennicutt and evans 2012 ; and the large body of literature therein ) . it is established that the sf law represents the universal scaling relation in galactic - scale star formation . detailed analyses of dependence of the law on the ism condition , particularly on the molecular gas have been recently obtained in several galaxies , and the molecular - gas regulated star formation scenario is proposed and modeled ( boissier et al . 2003 ; booth et al . 2007 ; leroy et al . 2008 ; krumholz et al . 2009 , 2013 ) . the sf law in the milky way has been investigated using individual star forming regions associated with molecular and hi clouds and spiral arms ( fuchs et al . 2006 ; luna et al . 2006 ; heiderman et al . 2010 ; willis et al . 2015 ; heyer et al . 2016 ; and the papers cited therein ) . recently , we analyzed the galactic distributions of hii regions , , and hi gases to obtain a global sf law in the entire milky way with spatial resolution of @xmath1 kpc in the galactic plane ( sofue and nakanishi 2016b ; paper i ) . we showed that hii regions are distributed in a star - forming disc with nearly constant full thickness of 92 pc in spatial coincidence with the molecular gas disc . with this thickness , the vertically averaged volume star formation rate ( sfr ) , @xmath2 , was shown to be related to the surface sfr , @xmath3 , by @xmath4 = \vsfr / [ \vsfrunitgy ] = 9.26 \times \ssfr / [ \sfu]$ ] . this relation will be used throughout this paper . in this paper , we examine the variations of both the volume and surface sf laws in the galaxy . the volume sf law in the whole galactic disc will be analyzed in detail for the first time , which can be obtained uniquely by analyzing the 3d data of hi and gas distributions . both the volume- and surface - sf laws will be shown to be not universal in the galaxy , but highly variable with the galacto - centric distance and dependent on the sfr itself . in the analyses we use the hii region catalogue in the milky way ( hou and han 2014 ) and the 3d maps of hi and gases ( nakanishi and sofue 2005 , 2006 , 2016 ; sofue and nakanishi 2016a ) , which were obtained by analyzing the leiden - argentine - bonn all sky hi survey ( kalberla et al . 2005 ) and galactic disc co line survey ( dame et al . 2001 ) . following the definition adopted in paper i , we express the volume sfr by power - law function of gas density , and represent it in logarithmic form as @xmath5 where @xmath6 is the vertically averaged volume sfr in the sf disc of full thickness 92 pc , and @xmath7 with @xmath8 being the coefficient between @xmath9 and @xmath10 defined in paper i. the coefficient @xmath11 represents the @xmath2 at @xmath12 . similarly , the surface sfr law is expressed by @xmath13 where the coefficient @xmath14 represents @xmath3 at @xmath15 . the sfr was calculated using the emission rates of uv photons of hii regions inferred from their excitation parameter observed by radio observations ( schraml and mezger 1969 ; hou and han 2014 ) . the incompleteness of mapping of hii regions was corrected for the detection probability , which is a function of the distance and intrinsic luminosity of each hii region . the distance errors are mostly kinematical distance errors due to intrinsic velocity dispersion of hii regions , while recombination line velocity measurements had better resolution . they are on the same order as those for the hi and gases , @xmath16 to 1 kpc , as described below . the number of effective independent data points in each plot of sf law is the number of @xmath17 kpc grids in the galactic plane , which is taken into account for calculating the standard errors of the plotted quantities . the map of sfr was obtained by smoothing the uv emission rates by a gaussian beam of fwhm 1 kpc ( @xmath18 kpc ) in the galactic plane . the vertical resolution depend on the telescope resolution and positioning , and was of the order of several arc minutes , or @xmath19 pc for @xmath20 at @xmath21 kpc distance , and was better than that for the gases as below . however , since the number of hii regions was not sufficient in order to divide the disc into a meaningful number of layers , we do not discuss the vertical variation . the distribution maps of hi and gases in the galaxy have spatial resolution depending on their errors in kinematical distances and therefore on the distance from the sun and direction . except for the velocity - degenerated regions in the galactic center and anti - center directions as well as along the solar circle , the distance error is about @xmath16 to 1 kpc for an effective radial velocity uncertainty corresponding to the interstellar velocity dispersion of @xmath22 ( see eq . 23 of sofue ( 2011 ) ) , although the observational resolution of the lab hi survey was @xmath23 . we examined how the sf relation varies with the galacto - centric distance @xmath0 . we obtained the same plots both for the volume and surface sf laws , and found the same or similar behaviors . first , we divided the galaxy into two areas of the inner disc at @xmath24 kpc , and outer disc at @xmath25 kpc , and derived the sf law in each area . figures [ rho8kpc ] and [ sig8kpc ] show the results , where the straight lines are the least - squares fitting to the plots in the log - log plane . table [ tab1 ] and [ tab2 ] list the best - fit values for the indices and coefficients . in the first row of the tables , we list the parameters for the whole galaxy obtained in paper i for comparison . there is remarkable difference in the sf laws in the inner and outer discs . the inner sf law shows significantly smaller power - law index compared to that in the outer disc . also , the average value of sfr is greater in the inner disc , trivially reflecting the fact that the sf activity is higher in the inner disc because of the higher gas density . in order to examine finer variations with @xmath0 , we further divide the galaxy into a larger number of rings of 2 kpc width . the volume- and surface sf laws made in individual rings are shown in figures [ rho2kpc ] and [ sig2kpc ] , respectively , in the appendix . we fitted each plot by the least - squares fitting , and show the result by inserted straight lines . the best - fit values are listed in tables [ tab1 ] and [ tab2 ] . there are shorter - scale variations of the sf law with the radius . besides the generally flat index in the inner galaxy , the flattest , or an even inverse index of the volume sf law was found at @xmath26 kpc . on the other hand , the surface sf law exhibits the flattest index at @xmath27 kpc , and the innermost and galactic center regions show again steep index . the inverse volume sf law at @xmath26 kpc indicates that the sfr decreases with increasing gas density . this apparently strange behavior could be understood , if some regulation mechanism is working in such a way that the sf is strongly suppressed , when the sfr exceeds a certain threshold value . in the present case , this threshold exists at @xmath28 . it must be also emphasized that the region coincides in position with the 4-kpc molecular ring with the highest sfr in the galaxy . another remarkable nature of the sf laws at @xmath29 kpc is the coincidence of both the index and coefficient for with those for hi and total gases . this may reflect the facts that the inner galactic disc is dominated by the molecular gas and that the hi gas density is saturated at a threshold density related to the molecular gas density ( sofue and nakanishi 2016a ) . in contrast , the sf law for gas in the outer galaxy at @xmath30 kpc is significantly displaced from those for hi and total gases , showing a much milder index . also , the outer sf laws are similar to that obtained in the entire galactic disc in paper i. this means that the global sf law is more determined by the laws in the outer galaxy for the larger area , and hence , for the larger weight in the averaging . thus , the global sf law , often regarded as the universal scaling relation , reflects more the sf relations in the outer disc . this sounds paradoxical in the sense that the universal sf law may not necessarily be the direct measure of the most active star formation in the galaxy . namely , the most active sf in the galaxy can not be investigated by the analysis of the global sf law . this paradox reminds us of the sf law in starburst galaxies , which is significantly displaced from the universal sf law for normal galaxies ( komugi et al . 2005 , 2006 ) . such displacement could be due to selectively observed active sf regions in starbursts , having intrinsically larger sf coefficients . [ cols="<,<,<,<,<,<,<",options="header " , ] [ tabemp ] the highest @xmath11 and @xmath14 , and hence the most active sf activity , is found in a radial zone at @xmath31 kpc centered by the 4-kpc molecular ring ( figures [ r - a - alpha ] , [ r - b - beta ] ) . the corresponding schmidt plots in the panels at @xmath32 kpc of figures [ rho2kpc ] and [ sig2kpc ] exhibit flat slopes with indices less than 1 . such a flat sf law may imply that the star formation in high - sfr regions tends to be saturated by molecular gas dissociation due to uv emission as well as by injection of kinetic and thermal energies by stellar winds and supernova explosions in active sf regions ( krumholz et al . the schmidt - kennicutt law between the volume sfr @xmath2 and volume gas densities @xmath33 of hi , , and total gases was shown to vary with the galacto - centric distance @xmath0 . it was shown that the sf power - law index @xmath34 is significantly smaller in the inner disc than in the outer galaxy . even an inverse index was found at radii near the 4-kpc molecular ring . the sf index was also found to be inversely correlated with the coefficient @xmath11 of sfr , suggesting regulation of sfr by increasing sf activity . similar results were obtained also for the surface sf laws . we , thus , suggest that the resolved sf law in the galaxy may be neither uniform , nor does give the universal scaling relation . variation of the sf index depending on the gaseous phase may be explained by the molecular gas regulated star formation ( e.g. , krumhotz et al . 2009 ) . if we assume that molecular clouds are formed from hi gas by a similar power law as @xmath35 , we may expect that @xmath36 , and so for @xmath37 . the variable sf index may be , thus , a measure of `` directness '' of star formation in the sense that the more additional process is involved prior to sf , the steeper becomes the index . this may also explain the steeper index of sf laws for surface gas densities than for volume densities . in other words , the @xmath34 index is a more direct measure of star formation than @xmath37 , trivially indicating that the star formation is more directly related to the volume density than to the column density . this consideration may also apply to the sf law for total gas , for which the index is a mixture of indices for and hi , and is steeper than the index for gas . it was also found in paper i that the sf power - law index for total gas has a bend from steep to milder around the threshold density of @xmath38 h @xmath39 , which we reproduce in figure [ trans ] . in the figure we compare this sf law with the plot of molecular gas density against the total gas density , where also a bend exists . both the bends appear at the same threshold density of @xmath40 . these bent behaviors are in agreement with the model of molecular - gas regulated star formation in galaxies ( booth et al . 2007 ; bigiel et al . 2008 ) , and the bent sf law readily found in galaxies ( krumholz et al . 2009 ) . ( raw data from paper i : grey circles ) . triangles show gas density ( @xmath41 plotted against @xmath42 , showing phase transition from hi to around the threshold density of @xmath40 h @xmath39 . the bend from steep to mild index occurs around this threshold density.,width=264 ] the hi - to- transition in low density regions , where star formation is settled from forming gas , makes the sf index steeper than that in the molecular - gas dominant regions . the steeper index in the outer galaxy with lower density and lower molecular fraction reflects such indirect star formation , where the index tends to become larger because of the additional processes of transition from hi to and formation of molecular clouds . the @xmath43 law is , thus , considered to be a direct measure of star formation ( krumhotz et al . the other relations for @xmath44 , and @xmath45 must be analyzed more carefully by taking into account the intervening processes such as the phase - transition between the and hi , and the formation and dissociation of molecular clouds . the current studies used to use the surface relations , which implicitly contain the indirect or unrelated hi gas by projecting on the galactic plane . + _ acknowledgements _ : the author thanks dr . h. nakanishi of the kagoshima university for the hi and 3d cube data . bigiel f. , leroy a. , walter f. , brinks e. , de blok w. j. g. , madore b. , thornley m. d. , 2008 , aj , 136 , 2846 booth c. m. , theuns t. , okamoto t. , 2007 , mnras , 376 , 1588 boissier s. , prantzos n. , boselli a. , gavazzi g. , 2003 , mnras , 346 , 1215 dame t. m. , hartmann d. , thaddeus p. , 2001 , apj , 547 , 792 fuchs b. , jahrei h. , flynn c. , 2009 , aj , 137 , 266 elmegreen , b. g. 1993 , apj , 411 , 170 heyer , m. , & dame , t. m. 2015 , araa , 53 , 583 heyer , m. , gutermuth , r. , urquhart , j. s. , et al . 2016 , aa , 588 , a29 heiderman a. , evans n. j. , ii , allen l. e. , huard t. , heyer m. , 2010 , apj , 723 , 1019 - 1037 hou l. g. , han j. l. , 2014 , a&a , 569 , a125 kalberla , p. m. w. , burton , w. b. , hartmann , d. , et al . 2005 , aa , 440 , 775 kennicutt r. c. 1998a , apj , 498 , 541 kennicutt , r. c. 1998b araa 36 , 189 . kennicutt , r. c. , & evans , n. j. 2012 , araa , 50 , 531 komugi s. , sofue y. , egusa f. , 2006 , pasj , 58 , 793 komugi s. , sofue y. , nakanishi h. , onodera s. , egusa f. , 2005 , pasj , 57 , 733 krumholz , m. r. , mckee , c. f. and tumlinson , j. 2009 , apj , 693 , 235 . lada c. j. , forbrich j. , lombardi m. , alves j. f. , 2012 , apj , 745 , 190 leroy a. k. , walter f. , brinks e. , bigiel f. , de blok w. j. g. , madore b. , thornley m. d. , 2008 , aj , 136 , 2782 leroy a. k. , et al . , 2013 , aj , 146 , 19 liu g. , koda j. , calzetti d. , fukuhara m. , momose r. , 2011 , apj , 735 , 63 luna a. , bronfman l. , carrasco l. , may j. , 2006 , apj , 641 , 938 nakanishi h. , sofue y. , 2003 , pasj , 55 , 191 nakanishi h. , sofue y. , 2006 , pasj , 58 , 847 sofue y. , nakanishi h. , 2016 , pasj , 68 , 63 schmidt m. , 1959 , apj , 129 , 243 schraml , j. , & mezger , p. g. 1969 , apj , 156 , 269 sofue y. , 2011 , pasj , 63 , 813 sofue y. , nakanishi h. , 2016a , pasj , 68 , 63 sofue , y. , nakanishi , h. 2016b , pasj in press ( arxiv:1610.05396 ) , paper i willis , s. , guzman , a. , marengo , m. , et al . 2015 , apj , 809 , 87 in this appendix we show volume - sf laws ( plots of @xmath2 against @xmath46 ) made in individual annulus rings of 2 kpc width in figure [ rho2kpc ] , and surface - sf laws ( @xmath3 against @xmath47 ) in figure [ sig2kpc ] . the inserted lines show the least - squares fitting results , and the parameters ( indices and coefficients ) are listed in tables [ tab1 ] and [ tab2 ] .	variation of the volume- and surface - schmidt laws ( star - formation or sf law ) with the galacto - centric distance @xmath0 was investigated using 3d distributions of hii regions , hi , and molecular ( ) gases in the milky way . both the power - law index and sf coefficient were found to be variable with @xmath0 . the index is flatter in the inner disc than in the outer galaxy , and the coefficient is larger in the inner disc , decreasing steeply outward . there is also a mutual anti - correlation between the index and sf coefficient , and the sf law can be expressed by a single - parameter function of the sf coefficient . the variable sf law is discussed in relation to the self - regulation star formation . galaxies : individual ( milky way ) ism : hi gas ism : molecular gas ism : hii regions ism : star formation
_ resource usage analysis _ infers the aggregation of some numerical properties , like memory usage , time spent in computation , or bytes sent over a wire , throughout the execution of a piece of code . such numerical properties are known as _ resources_. the expressions giving the usage of resources are usually given in terms of the sizes of some input arguments to procedures . our starting point is the methodology outlined by @xcite and @xcite , characterized by the setting up of recurrence equations . in that methodology , the size analysis is the first of several other analysis steps that include cardinality analysis ( that infers lower and upper bounds on the number of solutions computed by a predicate ) , and which ultimately obtain the resource usage bounds . one drawback of these proposals , as well as most of their subsequent derivatives , is that they are only able to cope with size information about subterms in a very limited way . this is an important limitation , which causes the analysis to infer trivial bounds for a large class of programs . for example , consider a predicate which computes the factorials of a list : p6.3cmp6 cm .... listfact ( [ ] , [ ] ) . listfact([e\|r],[f\|fr ] ) : - fact(e , f ) , listfact(r , fr ) . .... & .... fact(0,1 ) . fact(n , m ) : - n1 is n - 1 , fact(n1 , m1 ) , m is n * m1 . .... intuitively , the best bound for the running time of this program for a list @xmath0 is @xmath1 , where @xmath2 and @xmath3 are constants related to the unification and calling costs . but with no further information , the upper bound for the elements of @xmath0 must be @xmath4 to be on the safe side , and then the returned overall time bound must also be @xmath4 . in a previous paper @xcite we focused on a proposal to improve the size analysis based on _ sized types_. these sized types are similar to the ones present in @xcite for functional programs , but our proposal includes some enhancements to deal with regular types in logic programs , developing solutions to deal with the additional features of logic programming such as non - determinism and backtracking . while in that paper we already hinted at the fact that the application of our sized types in resource analysis could result in considerable improvement , no description was provided of the actual resource analysis . this paper is complementary and fills this gap by describing a new resource usage analysis with two novel aspects . firstly , it can _ take advantage of the new information contained in sized types_. furthermore , this resource analysis is _ fully based on abstract interpretation _ , i.e. , not just the auxiliary analyses but also the resource analysis itself . this allows us to integrate resource analysis within the plai abstract interpretation framework @xcite in the ciaopp system , which brings in features such as _ multivariance _ , fixpoints , and assertion - based verification and user interaction for free . we also perform a performance assessment of the resulting global system . in section [ sec : overview ] we give a high - level view of the approach . in the following section we review the abstract interpretation approach to size analysis using sized types . section [ sec : resources ] gets deeper into the resource usage analysis , our main contribution . experimental results are shown in section [ sec : results ] . finally we review some related work and discuss future directions of our resource analysis work . we give now an overview of our approach to resource usage analysis , and present the main ideas in our proposal using the classical ` append/3 ` predicate as a running example : .... append ( [ ] , s , s ) . append([e\|r ] , s , [ e\|t ] ) : - append(r , s , t ) . .... the process starts by performing the regular type analysis present in the ciaopp system @xcite . in our example , the system infers that for any call to the predicate ` append(x , y , z ) ` with ` x ` and ` y ` bound to lists of numbers and ` z ` a free variable , if the call succeeds , then ` z ` also gets bound to a list of numbers . the set of `` list of numbers '' is represented by the regular type `` listnum , '' defined as follows : .... listnum - > [ ] \| .(num , listnum ) .... from this regular type definition , sized type schemas are derived . in our case , the sized type schema @xmath5 is derived from @xmath6 . this schema corresponds to a list that contains a number of elements between @xmath2 and @xmath3 , and each element is between the bounds @xmath7 and @xmath8 . it is defined as : @xmath9 from now on , in the examples we will use @xmath10 and @xmath11 instead of @xmath6 and @xmath12 for the sake of conciseness . the next phase involves relating the sized types of the different arguments to the ` append/3 ` predicate using recurrence ( in)equations . let @xmath13 denote the sized type schema corresponding to argument ` x ` in a call ` append(x , y , z ) ` ( created from the regular type inferred by a previous analysis ) . we have that @xmath13 denotes @xmath14 . similarly , the sized type schema for the output argument ` z ` is @xmath15 , denoted by @xmath16 . now , we are interested in expressing bounds on the length of the output list ` z ` and the value of its elements as a function of size bounds for the input lists ` x ` and ` y ` ( and their elements ) . for this purpose , we set up a system of inequations . for instance , the inequations that are set up to express a lower bound on the length of the output argument ` z ` , denoted @xmath17 , as a function on the size bounds of the input arguments ` x ` and ` y ` , and their subarguments ( @xmath18 , and @xmath19 ) are : @xmath20 note that in the recurrence inequation set up for the second clause of ` append/3 ` , the expression @xmath21 ( respectively @xmath22 ) represents the size relationship that a lower ( respectively upper ) bound on the length of the list in the first argument of the recursive call to ` append/3 ` is one unit less than the length of the first argument in the clause head . as the number of size variables grows , the set of inequations becomes too large . thus , we propose a compact representation . the first change in our proposal is to write the parameters to size functions directly as sized types . now , the parameters to the @xmath17 function are the sized type schemas corresponding to the arguments ` x ` and ` y ` of the ` append/3 ` predicate : @xmath23 in a second step , we group together all the inequalities of a single sized type . as we always alternate lower and upper bounds , it is always possible to distinguish the type of each inequality . we do not write equalities , so that we do not use the symbol @xmath24 . however , we always write inequalities of both signs ( @xmath25 and @xmath26 ) for each size function , since we compute both lower and upper size bounds . thus , we use a compact representation @xmath27 for the symbols @xmath25 and @xmath26 that are always paired . for example , the expression : @xmath28 represents the conjunction of the following size constraints : @xmath29 after setting up the corresponding system of inequations for the output argument ` z ` of ` append/3 ` , and solving it , we obtain the following expression : @xmath30 that represents , among others , the relation @xmath31 ( resp . @xmath32 ) , expressing that a lower ( resp . upper ) bound on the length of the output list ` z ` , denoted @xmath33 ( resp . @xmath34 ) , is the addition of the lower ( resp . upper ) bounds on the lengths of ` x ` and ` y ` . it also represents the relation @xmath35 ( resp . @xmath36 ) , which expresses that a lower ( resp . upper ) bound on the size of the elements of the list ` z ` , denoted @xmath37 ( resp . @xmath38 ) , is the minimum ( resp . maximum ) of the lower ( resp . upper ) bounds on the sizes of the elements of the input lists ` x ` and ` y ` . resource analysis builds upon the sized type analysis and adds recurrence equations for each resource we want to analyze . apart from that , when considering logic programs , we have to take into account that they can fail or have multiple solutions when executed , so we need an auxiliary _ cardinality analysis _ to get correct results . let us focus now on cardinality analysis . let @xmath39 and @xmath40 denote lower and upper bounds on the number of solutions respectively that predicate ` append/3 ` can generate . following the program structure we can infer that : @xmath41 @xmath42 the solution to these inequations is @xmath43 , so we have inferred that ` append/3 ` generates at least ( and at most ) one solution . thus , it behaves like a function . when setting up the equations , we have used our knowledge that ` append/3 ` can not fail when given lists as arguments . if not , the lower bound in the number of solutions would be 0 . now we move forward to analyzing the number of resolution steps performed by a call to ` append/3 ` ( we will only focus on upper bounds , @xmath44 , for brevity ) . for the first clause , we know that only one resolution step is needed , so : @xmath45 the second clause performs one resolution step plus all the resolution steps performed by all possible backtrackings over the call in the body of the clause . this number of possible backtrackings is bounded by the number of solutions of the predicate . so the equation reads : @xmath46 solving these equations we infer that an upper bound on the number of resolution steps is the ( upper bound on the length ) of the input list ` x ` plus one . this is expressed as : @xmath47 as shown in the ` append ` example , the ( bound ) variables that we relate in our inequations come from sized types , which are ultimately derived from the regular types previously inferred for the program . among several representations of regular types used in the literature , we use one based on _ regular term grammars _ , equivalent to @xcite but with some adaptations . a _ type term _ is either a _ base type _ @xmath48 ( taken from a finite set ) , a _ type symbol _ @xmath49 ( taken from an infinite set ) , or a term of the form @xmath50 , where @xmath51 is a @xmath11-ary function symbol ( taken from an infinite set ) and @xmath52 are _ type terms_. a _ type rule _ has the form @xmath53 , where @xmath54 is a _ type symbol _ and @xmath55 a _ type term_. a _ regular term grammar _ @xmath56 is a set of _ type rules_. to devise the abstract domain we focus specifically on the generic and - or trees procedure of @xcite , with the optimizations of @xcite . this procedure is _ generic _ and goal dependent : it takes as input a pair @xmath57 representing a predicate along with an abstraction of the call patterns ( in the chosen _ abstract domain _ ) and produces an abstraction @xmath58 which overapproximates the possible outputs . this procedure is the basis of the plai abstract analyzer present in ciaopp @xcite , where we have integrated an implementation of the proposed size analysis . the formal concept of _ sized type _ is an abstraction of a set of herbrand terms which are a subset of some regular type @xmath54 and meet some lower- and upper - bound size constraints on the number of _ type rule applications_. a grammar for the new sized types follows : ' '' '' [ cols=">,^ , < , > " , ] we have constructed a prototype implementation in ciao by defining the abstract operations for sized type and resource analysis that we have described in this paper and plugging them into ciaopp s plai implementation . our objective is to assess the gains in precision in resource consumption analysis . table [ expresults ] shows the results of the comparison between the new lower ( * _ lb _ * ) and upper bound ( * _ ub _ * ) resource analyses implemented in ciaopp , which also use the new size analysis ( columns _ new _ ) , and the previous resource analyses in ciaopp @xcite ( columns _ previous _ ) . we also compare ( for upper bounds ) with _ raml _ s analysis @xcite ( column _ raml _ ) . although the new resource analysis and the previous one infer concrete resource usage bound functions ( as the ones in @xcite ) , for the sake of conciseness and to make the comparison with raml meaningful , table [ expresults ] only shows the complexity orders of such functions , e.g. , if the analysis infers the resource usage bound function @xmath59 , and @xmath60 , table [ expresults ] shows @xmath61 . the parameters of such functions are ( lower or upper ) bounds on input data sizes . the symbols used to name such parameters have been chosen assuming that lists of numbers @xmath62 have size @xmath63 , lists of lists of lists of numbers have size @xmath64 , and numbers have size @xmath65 . table [ expresults ] also includes columns with symbols summarizing whether the new ciaoppresource analysis improves on the previous one and/or _ raml _ s : @xmath66 ( resp . @xmath67 ) indicates more ( resp . less ) precise bounds , and @xmath24 the same bounds . the new size analysis improves on ciaopp s previous resource analysis in most cases . moreover , raml can only infer polynomial costs , while our approach is able to infer other types of cost functions , as is shown for the divide - and - conquer benchmarks ` hanoi ` and ` fib ` , which represent a large and common class of programs . for predicates with polynomial cost , we get equal or better results than raml . several other analyses for resources have been proposed in the literature . some of them just focus on one particular resource ( usually execution time or heap consumption ) , but it seems clear that those analyses could be generalized . we already mentioned raml @xcite in section [ sec : results ] . their approach differs from ours in the theoretical framework being used : raml uses a type and effect system , whereas our system uses abstract interpretation . another important difference is the use of polynomials in raml , which allows a complete method of resolution but limits the type of closed forms that can be analyzed . in contrast , we use recurrence equations , which have no complete decision procedure , but encompass a much larger class of functions . type systems are also used to guide inference in @xcite and @xcite . in @xcite , the authors use sparsity information to infer asymptotic complexities . in contrast , we only get closed forms . similarly to ciaopp s previous analysis , the approach of @xcite applies the recurrence equation method directly ( i.e. , not within an abstract interpretation framework ) . @xcite shows a complexity analysis based on abstract interpretation over a step - counting version of functional programs . @xcite uses symbolic evaluation graphs to derive termination and complexity properties of logic programs . in this paper we have presented a new formulation of resource analysis as a domain within abstract interpretation and which uses as input information the sized types that we developed in @xcite . we have seen how this approach offers benefits both in the quality of the bounds inferred by the analysis , and in the ease of implementation and integration within a framework such as plai / ciaopp . in the future , we would like to study the generalization of this framework to different behaviors regarding aggregation . for example , when running tasks in parallel , the total time is basically the maximum of both tasks , but memory usage is bounded by the sum of them . another future direction is the use of more ancillary analyses to obtain more precise results . also , since we use sized types as a basis , any new research that improves such analysis will directly benefit the resource analysis . finally , another planned enhancement is the use of mutual exclusion analysis ( already present in ciaopp ) to aggregate recurrence equations in a better way . e. albert , s. genaim , and a. n. masud . ore precise yet widely applicable cost analysis . in r. jhala and d. schmidt , editors , _ 12th verification , model checking , and abstract interpretation ( vmcai11 ) _ , volume 6538 of _ lecture notes in computer science _ , pages 3853 . springer verlag , january 2011 . r. bagnara , a. pescetti , a. zaccagnini , and e. zaffanella . : towards computer algebra support for fully automatic worst - case complexity analysis . technical report , 2005 . available from http://arxiv.org/. f. bueno , p. lpez - garca , and m. hermenegildo . ultivariant non - failure analysis via standard abstract interpretation . in _ 7th international symposium on functional and logic programming ( flops 2004 ) _ , number 2998 in lncs , pages 100116 , heidelberg , germany , april 2004 . springer - verlag . s. k. debray , n .- w . lin , and m. hermenegildo . ask granularity analysis in logic programs . in _ proc . of the 1990 acm conf . on programming language design and implementation _ , pages 174188 . acm press , june 1990 . s. k. debray , p. lpez - garca , m. hermenegildo , and n .- w . ower bound cost estimation for logic programs . in _ 1997 international logic programming symposium _ , pages 291305 . mit press , cambridge , ma , october 1997 . jrgen giesl , thomas strder , peter schneider - kamp , fabian emmes , and carsten fuhs . symbolic evaluation graphs and term rewriting : a general methodology for analyzing logic programs . in _ ppdp _ , pages 112 . acm , 2012 . p. lpez - garca , l. darmawan , and f. bueno . ramework for verification and debugging of resource usage properties . in m. hermenegildo and t. schaub , editors , _ technical communications of the 26th intl . conference on logic programming ( iclp10 ) _ , volume 7 of _ leibniz international proceedings in informatics ( lipics ) _ , pages 104113 , dagstuhl , germany , july 2010 . leibniz - zentrum fuer informatik . j. navas , e. mera , p. lpez - garca , and m. hermenegildo . ser - definable resource bounds analysis for logic programs . in _ 23rd international conference on logic programming ( iclp07 ) _ , volume 4670 of _ lecture notes in computer science_. springer , 2007 . g. puebla and m. hermenegildo . ptimized algorithms for the incremental analysis of logic programs . in _ international static analysis symposium ( sas 1996 ) _ , number 1145 in lncs , pages 270284 . springer - verlag , september 1996 . pedro b. vasconcelos and kevin hammond . inferring cost equations for recursive , polymorphic and higher - order functional programs . in philip w. trinder , greg michaelson , and ricardo pena , editors , _ ifl _ , volume 3145 of _ lecture notes in computer science _ , pages 86101 . springer , 2003 . c. vaucheret and f. bueno . ore precise yet efficient type inference for logic programs . in _ international static analysis symposium _ , volume 2477 of _ lecture notes in computer science _ , pages 102116 . springer - verlag , september 2002 .	we present a novel general resource analysis for logic programs based on sized types.sized types are representations that incorporate structural ( shape ) information and allow expressing both lower and upper bounds on the size of a set of terms and their subterms at any position and depth . they also allow relating the sizes of terms and subterms occurring at different argument positions in logic predicates . using these sized types , the resource analysis can infer both lower and upper bounds on the resources used by all the procedures in a program as functions on input term ( and subterm ) sizes , overcoming limitations of existing analyses and enhancing their precision . our new resource analysis has been developed within the abstract interpretation framework , as an extension of the sized types abstract domain , and has been integrated into the ciao preprocessor , ciaopp . the abstract domain operations are integrated with the setting up and solving of recurrence equations for both , inferring size and resource usage functions . we show that the analysis is an improvement over the previous resource analysis present in ciaopp and compares well in power to state of the art systems . [ firstpage ]
in this paper , we describe the results of a centimetre - wave radio study of iras 20100@xmath04156 , an ultra - luminous infrared galaxy ( ulirg ) with a redshift @xmath3 = 0.129 @xcite . iras 20100@xmath04156 hosts a very luminous megamaser ( actually a ` gigamaser ' ) in the @xmath4 , j = @xmath5 , f = 2 @xmath0 2 ( 1667.3590 mhz ) main line transition of the hydroxyl ( oh ) molecule . oh megamasers are thought to occur when radio continuum emission from the galaxy is amplified by oh molecules whose energy - level population has been inverted by a far - infrared radiation field generated by dust ( see * ? ? ? * for an overview ) . optical and infrared images reveal this to be part of a merging pair or possibly a trio of galaxies accompanied by tidal tails and clumps of star - forming gas and dust . both optical and infrared data indicate that the system is in a starburst phase @xcite , has a low ionisation nuclear emission region ( liner)-like spectrum @xcite and a star formation rate , estimated from far - infrared emission , of approximately 700 solar masses per year @xcite . searches for x - rays and polarised h@xmath6 emission revealed no definitive evidence of an active galactic nucleus ( agn ) @xcite , however a _ spitzer _ space telescope study using a diagnostic based upon the 5 to 8 @xmath7 m brightness ratio found evidence for a 20 per cent bolometric contribution from an agn @xcite . this agn is presumably deeply embedded within a dusty starburst region at the centre of iras 20100@xmath04156 . the oh megamaser in iras 20100@xmath04156 was discovered using the parkes radio telescope by @xcite and a slightly improved spectrum was later published by @xcite . the former paper identified a double peak at 38,697 and 38,648 km s@xmath1 in the barycentric rest frame with an oh peak flux density of 200 mjy and a broad ` pedestal ' of emission . iras 20100@xmath04156 was subsequently observed using the australia telescope compact array ( atca ) as part of a study of magnetic fields in oh megamaser galaxies published by @xcite . they observed the line as having slightly lower peak flux density ( @xmath2165 mjy ) and found a previously unidentified ( 73 mjy ) peak at approximately 38,500 km s@xmath1 , although an absolute flux calibration was made difficult by the lack of a robust understanding of the spectral baseline due to a narrow observing bandwidth . in this study , we provide a pair of independent observations of the oh megamaser in iras 20100@xmath04156 in a third observing epoch . the observations were taken as part of spectral line commissioning of the australian square kilometre array pathfinder ( askap ) using the boolardy engineering test array ( hereafter beta ; * ? ? ? * ) and a follow - up observation was made using the atca . the aims of the study were to ( i ) test the spectral line capabilities of beta , ( ii ) compare the results to those obtained with the atca and ( iii ) investigate the generation mechanism , variability and kinematics of the oh megamaser emission in iras 20100@xmath04156 . beta comprises six 12-metre diameter parabolic reflectors outfitted with prototype askap phased array feeds @xcite based on a connected - element chequerboard " array @xcite . we used five antennas of beta ; antennas 1 ( diggidumble ) , 6 ( biyarli ) , 8 ( bimba ) , 9 ( gagurla ) and 15 ( minda ) , to observe the galaxy iras 20100@xmath04156 ( right ascension 20@xmath813@xmath929@xmath10509 , declination -41@xmath1147@xmath1235@xmath1348 , j2000 ) from ut 20:00 march 3rd @xmath0 08:00 march 4th 2015 ( scheduling block i d 1496 ) and from ut 06:25 @xmath0 10:42 on 12th may 2015 ( scheduling block i d 1791 ) . we also observed the radio galaxy pks b1934 - 638 for 15 minutes on each day to provide calibration data before the target was observed . the correlator was configured to record a total bandwidth of 304 mhz between 1399.5 and 1703.5 mhz in 16,416 spectral channels , each with a width of approximately 18.5 khz ( @xmath14@xmath15 = 3.76 km s@xmath1 at 1477 mhz ) . nine beams were generated , oriented in a 3 @xmath16 3 square pattern with an inter - beam spacing of 0.66 degrees , with the central beam pointed directly at the target source . we subsequently obtained director s discretionary time on the atca to provide an independent verification of our results . the observations ( observing code cx322 ) were conducted in 6 km array configuration ( 6a ) between ut 23:45 16th april @xmath0 00:55 17th april 2015 . the correlator @xcite was configured to simultaneously record a 2 ghz - wide continuum band with 32 @xmath16 64 mhz channels ( velocity resolution 11507.251 km s@xmath1 ) and four overlapping zoom modes , centred at 1476 mhz . the output spectral line data set had 5120 spectral channels with a spectral resolution of about 31.3 khz ( or 5.619 km s@xmath1 ) . we again used pks b1934 - 638 as a primary calibrator source . observations of the target source ( 25 minutes ) were interspersed with three short ( 2 minute ) observations of the phase calibrator pks 1954 - 388 . the total integration time on the target source was 56 minutes , although a small amount of data was flagged due to a suspected phase jump on antenna 3 . our beta data were flagged , imaged and calibrated following the method described by @xcite , although our procedure here deviates slightly . notable changes include an increase in the spectral resolution of the full - band coarse - channel data to 1mhz , used for self - calibration . this was done to reduce the effect of radio frequency interference ( rfi ) , which is more prevalent in this frequency band than at frequencies between 700 mhz and 1ghz . the flux scale was corrected using a model of pksb1934 - 638 @xcite , and we applied a primary beam correction assuming a simple gaussian model , scaled with frequency . the beta flux scale was verified against the sydney university molonglo sky survey source catalogue @xcite by @xcite and against the nrao vla sky survey catalogue @xcite by @xcite . since our target source is at the centre of the beam , we expect any systematic deviation in the flux density over the period of our observation to be dominated by variation in the gain from our observation of pksb1934 - 638 . we estimate the uncertainty in the beta flux scale to be approximately 7% . this was determined by adding in quadrature the 3% uncertainty introduced by bootstrapping the flux to pksb1934 - 638 ; a stochastic component of 4% determined by comparing the flux densities of several continuum point - sources close to the field centre over the two beta observing epochs ; and an approximately 5% component introduced by fitting and removing a baseline ripple introduced by nearby rfi . in fig . [ figure1 ] we show the resultant spectrum at the observed frequency of the red - shifted 18-cm oh line . the atca zoom mode ( 31.3 khz spectral resolution ) data were processed as follows . the raw uvdata were loaded into miriad @xcite and some edge channels and those known to contain bad data or correlator artefacts were flagged using atlod . visibilities with extreme amplitudes ( @xmath17500 jy ) were flagged using uvflag and further interactive flagging was carried out using blflag . bandpass and polarisation leakage corrections were calculated for the primary flux calibrator pks b1934 - 638 using mfcal and gpcal , respectively . bandpass corrections were copied to the phase calibrator pks 1954 - 388 before polarisation leakage corrections were determined using that source and the gpcal task . using gpboot , the absolute flux scale of the phase calibrator was tied to pks b1934 - 638 , using the same method as for the beta observations . the uncertainty in the absolute flux scale using this model was estimated by @xcite to be 1@xmath02% . adding this in quadrature with the 3% systematic uncertainty introduced by bootstrapping the flux to pksb1934 - 638 gives an overall uncertainty in the atca flux density scale of 4% . we then adjusted the labelling of velocity channels in the target source data for the doppler correction during the observation using uvaver and we converted the data to the barycentric rest frame . the calibration solutions from pks 1954 - 388 were copied to the target source iras 20100@xmath04156 . a baseline ripple was reduced by fitting a fifth - order polynomial to the line - free channels surrounding the oh maser and subtracting this using uvlin . the task invert was used to generate a calibrated image cube of iras 20100@xmath04156 . the peak pixel in the cube was located using maxfit and an average spectrum was extracted at that peak position using uvspec . the beta spectrum shows significant spectral line emission in iras 20100@xmath04156 lying close to the redshift - corrected rest frequency of the f=2@xmath02 oh ( 1667 mhz ) molecular transition . the median 1@xmath18 root mean square ( rms ) noise in the beta spectrum ( determined using the full 304 mhz band ) is 16.4 mjy per beam per spectral channel . [ figure1 ] shows the observed spectrum , smoothed using third order hanning smoothing . a simple baseline subtraction was carried out , however the left - hand side of the spectrum is still slightly affected by a baseline ripple caused by strong broadband radio emission from a satellite at approximately 1485 mhz . the main peak ( a ) apparently has a double structure , however the level of noise in the beta spectrum means that our gaussian fitting method recognises only a single component . 4156 , showing the three gaussian fits and their sum superimposed on the measured signal from beta . we have labeled the narrow - line peaks a and b for reference . peak a has a double structure , however the level of noise in the beta spectrum means that our gaussian fitting method does not recognise the double structure as a significant.,width=302 ] we modelled the spectrum as the sum of three gaussian components using magicplot . the model was determined by increasing the number of gaussians until the residuals within the frequency range of the spectral line resembled the noise . magicplot employs a levenberg - marquardt nonlinear least squares curve fitting algorithm to minimise the residual sum of squares ( @xmath19 ) by iteratively varying the parameters of the input functions . table 1 lists the characteristic velocities , widths and amplitudes of the three gaussian components determined by the fitting routine and fig . [ figure1 ] shows the derived gaussian fit curves and their sum , plotted with the measured data . our modelling reveals a broad component of oh emission , superimposed with two narrower peaks at approximately 38,665 and 38,492 km s@xmath1 in the barycentric rest frame . the physical interpretation of the line structure is discussed in section [ discussion ] . we find no statistically significant emission at the redshift - corrected rest frequency of the oh f=1@xmath01 ( 1665-mhz ) line . the f=2@xmath01 ( 1720-mhz ) oh satellite line is a non - detection , which is unsurprising since the oh satellite lines are very rarely seen in megamaser galaxies @xcite . we were not able to observe the f=1@xmath02 ( 1612-mhz ) oh satellite line in these observations , as the redshifted rest frequency was outside our observed frequency range . the continuum peak flux density of the source is @xmath20 = [email protected] mjy , which was determined from an image made from the full 304 mhz frequency band centred at 1551.5 mhz observed with beta . this is consistent with published measurements of @xmath20 = [email protected] mjy at 843 mhz from the sydney university molonglo sky survey @xcite , @xmath20 = [email protected] mjy at 1.49 ghz @xcite and @xmath20 = 9.6@xmath213 mjy at 5 ghz @xcite using parkes , assuming a spectral index of @xmath6 = @xmath00.5 , which is a typical value for ulirgs at these frequencies . [ cols="^,^,^,^,^,^,^ " , ] the flux density of the parkes continuum source iras 20100@xmath04156 was quoted as 9.6@xmath213 mjy @xcite . this very large ( 30% ) uncertainty in the parkes flux scale could potentially account for the apparent drop in the oh flux density between the 1988 epoch and subsequent observations listed in table [ table3 ] . instrumental changes and differences in absolute flux scales notwithstanding , the changes in the oh line profile shape point to genuine variability in the narrow - line masers , which may be due to ( a ) interstellar scintillation or ( b ) substantive changes in the physical conditions of the maser environment . this could arise from stochastic changes in the maser gain path length caused by turbulence , or a relative movement of maser clumps relative to the line - of - sight . alternatively it could be related to a genuine reduction in the background source of radio continuum photons . previous studies of oh megamasers have in some cases found their narrow - line emission to be variable . @xcite found a dramatic change in some , but not all , spectral features from iras 21272 + 2514 , which the authors argued were likely to be very compact maser regions ( smaller than 2 pc ) that were undergoing interstellar scintillation . a dedicated monitoring program carried out with askap or meerkat @xcite would be able to to distinguish between these scenarios and disentangle systematic measurement and calibration effects for iras 20100@xmath04156 . this and previous studies ( e.g. * ? ? ? * ) have demonstrated that the murchison radio - astronomy observatory has a relatively clean spectral environment , particularly in the frequency range from 700 mhz to 1 ghz . the maximum redshift probed by the arecibo oh megamaser survey @xcite was effectively set by the rfi environment at arecibo , with an effective lower limit on the observing frequency of 1200 mhz . the spectral environment was also mentioned as a significant constraint in the green bank telescope oh megamaser survey @xcite . as terrestrial rfi impacts radio observatory sites around the world , askap s location and radio - quiet protection @xcite provide the opportunity to search for oh megamasers in the largely unexplored frequency range below 1200 mhz corresponding to @xmath22 . this work is very encouraging for upcoming searches for megamasers in the deep investigation of neutral gas origins ( dingo ; * ? ? ? * ) and the first large absorption line survey of h i ( flash ; sadler et al . 2016 , _ in prep . _ ) with askap . we have observed the oh megamaser in iras 20100@xmath04156 using beta and the atca . the oh maser peak flux density is approximately 100 mjy beam@xmath1 and the continuum flux density at 1.4 ghz ( using beta ) is 14.4 mjy . we compared the peak flux density , peak velocity and fwhm of the four major spectral features , finding that the parameters agree within the uncertainties of our gaussian model fitting . this gives us an independent verification of the scientific performance of askap s first - generation phased array feeds . our atca spectrum reveals at least two narrow - line spectral features that have not previously been reported . these lie at @xmath0409 and @xmath0562 km s@xmath1 to the blue - shifted side of the systemic velocity . there is also a possible pair of weak features on the red - shifted side of the spectrum that warrant further investigation , particularly since one of these lies very close to the redshift - corrected velocity of the oh f=@xmath23 ( 1665 mhz ) transition . we considered whether the high velocity masers originate from a compact circumnuclear ring or from the agn - driven molecular outflow in iras 20100@xmath04156 . high - resolution imaging will solve this puzzle , however we note that a 50 pc circumnuclear ring would suggest the presence of an enclosed mass of 3.8@xmath24 . we studied the variation of the oh spectrum between 1988 and 2015 , finding that the narrow spectral features have varied significantly in that time . this may be caused by interstellar scintillation or by intrinsic changes in the maser environment . a dedicated monitoring program carried out with askap or meerkat will be useful to monitor iras 20100@xmath04156 and understand more about the variation in its oh maser emission . sensitive high - resolution observations of these masers using vlbi will enable us to locate the compact maser spots , confirm the extent of the circumnuclear masers and better constrain the black hole mass . the australian ska pathfinder is part of the australia telescope national facility which is managed by csiro . operation of askap is funded by the australian government with support from the national collaborative research infrastructure strategy . establishment of the murchison radio - astronomy observatory was funded by the australian government and the government of western australia . askap uses advanced supercomputing resources at the pawsey supercomputing centre . we acknowledge the wajarri yamatji people as the traditional owners of the observatory site . parts of this research were conducted by the australian research council centre of excellence for all - sky astrophysics ( caastro ) , through project number ce110001020 . the authors thank the referee , andrew walsh , for his insightful comments on the manuscript . roy , a. , burton , m.j . , argo , m. , baan , w. , cunningham , m. , ellingsen , s. , green , j. , henckel , c. , hopkins , a. ivison , r. , jones , p. , kloekner , h.r . , staveley - smith , l. , norris , r.p . , robishaw , t. , urquhart , j. , walsh , a. , 2009 , askap survey science project proposal schinckel , a. , bunton , j. , chippendale , a. , gough , r. , hampson , g. , hay , s. , jackson , c. , jeganathan , k. , osullivan , j. , reynolds , j. , shaw , r. , wilson , c. 2011 , in : asia - pacific microwave conference 2011 , melbourne , australia , 5 - 8 december , 2011 wilson , carol ; storey , michelle and tzioumis , tasso , 2013 , in : asia - pacific international symposium and exhibition on electromagnetic compatibility : apemc 2013 . barton , a.c.t . : engineers australia , 2013 : 376 - 379	we report the discovery of new , high - velocity narrow - line components of the oh megamaser in iras 20100@xmath04156 . results from the australian square kilometre array pathfinder ( askap ) s boolardy engineering test array ( beta ) and the australia telescope compact array ( atca ) provide two independent measurements of the oh megamaser spectrum . we found evidence for oh megamaser clumps at @xmath0409 and @xmath0562 km s@xmath1 ( blue - shifted ) from the systemic velocity of the galaxy , in addition to the lines previously known . the presence of such high velocities in the molecular emission from iras 20100@xmath04156 could be explained by a @xmath250 pc molecular ring enclosing a @xmath23.8 billion solar mass black hole . we also discuss two alternatives , i.e. that the narrow - line masers are dynamically coupled to the wind driven by the active galactic nucleus or they are associated with two separate galactic nuclei . the comparison between the beta and atca spectra provides another scientific verification of askap s beta . our data , combined with previous measurements of the source enabled us to study the variability of the source over a twenty - six year period . the flux density of the brightest oh maser components has reduced by more than a factor of two between 1988 and 2015 , whereas a secondary narrow - line component has more than doubled in the same time . plans for high - resolution vlbi follow - up of this source are discussed , as are prospects for discovering new oh megamasers during the askap early science program . [ firstpage ] masers radio lines : galaxies galaxies : kinematics and dynamics galaxies : interactions quasars : supermassive black holes .
in this paper we develop a method to approximate steady - state solutions of the following one - dimensional thermistor problem : @xmath0 subject to boundary and initial conditions , @xmath1 @xmath2 and coupled with the electric potential equation : @xmath3 @xmath4 @xmath5 the motivation for studying this kind of problem is that - has important implications for a variety of technological processes . for example , it arises in the analytical study of phenomena associated with the occurrence of shear band in metal being deformed at high strain rates @xcite ; in the theory of gravitational equilibrium of polytropic stars @xcite ; in the investigation of the fully turbulent behavior of flows @xcite ; in modelling aggregation of cells via interaction with a chemical substance ( chemotaxis ) @xcite ; and specially in modelling electrical heating in a conductor @xcite . in this case , @xmath6 is the temperature of the conductor , @xmath7 is the electrical potential . functions @xmath8 and @xmath9 are , respectively , the electrical and thermal conductivities ; @xmath10 is the heat transfer coefficient . the condition is a condition of robin - type . when @xmath11 it is called an adiabatic condition . equation consists in the heat equation with joule heating as a source ; describes conservation of current in the conductor . the thermistor problem has been extensively studied by several authors @xcite , where existence and uniqueness of solutions are given . theoretical analysis , consisting in existence of solutions with the required regularity and which ensure error estimates of optimal order of convergence , are done in @xcite . to construct a numerical approximation of the steady state solution we use a numerical method to approximate the solution of the parabolic problem . this approach has been used by @xcite in the one - dimensional thermistor problem . further , in these last works authors consider the thermal conductivity @xmath12 equal to @xmath13 and a particular electrical conductivity , then they obtain the exact solution @xmath14 of the conservation problem - and so system - of thermistor problem is reduced to the following single heat conduction problem : @xmath15 subject to the boundary conditions - . in this paper , we propose to solve both equations and at the same time by using a finite element method and a fully crank - nicolson approach . the formulation of the finite element method is standard and is based on a variational formulation of the continuous problem . in section [ sec2 ] we give the variational formulation of problem - . an algorithm for solving the problem is then proposed in section [ sec3 ] . in section [ sec4 ] , numerical results are obtained for an appropriate test - problem . we divide the interval @xmath16 $ ] into @xmath17 equal finite elements @xmath18 . let @xmath19 be a partition of @xmath20 and @xmath21 the step length . by @xmath22 we denote a basis of the usual pyramid functions : @xmath23,\\ - \frac{1}{h}x+(1+j ) & \mbox { on } [ x_{j } , x_{j+1}],\\ 0 & otherwise . \end{cases}\ ] ] as indicated above , it is convenient to proceed in two steps with the derivation and analysis of the approximate solution of - . first , we write the problem in weak or variational form . we multiply the parabolic equation by @xmath24 ( for @xmath25 fixed ) , integrate over @xmath26 , and apply green s formula on the left - hand side , to obtain @xmath27 using the boundary condition we get @xmath28 we also have @xmath29 we now turn our attention to solve this system by discretization with respect to the time variable . we introduce a time step @xmath30 and time levels @xmath31 , where @xmath32 is a nonnegative integer , and denote by @xmath33 the approximation of @xmath34 to be determined . we use the backward euler galerkin method , which is defined by replacing the time derivative in by a backward difference @xmath35 . so the approximations @xmath36 , @xmath37 admit unique representations @xmath38 where @xmath39 , @xmath40 are unknown real coefficients to be determined . then , after decoupling , we have that @xmath41 for scheme , we have @xmath42 on the other hand , we have @xmath43 using boundary conditions and initial condition , it follows that @xmath44 then , we have the resulting system of equations : for @xmath45 , @xmath46 for @xmath47 , @xmath48 for @xmath49 , @xmath50 coming back to , the following may be stated in terms of the functions @xmath51 : find the coefficients @xmath39 in @xmath52 such that @xmath53 in matrix notation , this may be expressed as @xmath54 where @xmath55 @xmath56 and @xmath57 since the matrix @xmath58 and @xmath59 are gram matrices , in particular they are positive definite and invertible . thus , the above system of ordinary differential equations has obviously a unique solution . we solve the system for each time level . estimating each term of separately , we have : @xmath60 using the expression of @xmath61 and @xmath62 , we obtain @xmath63 in the same way , we have @xmath64 on other hand , we similarly have @xmath65 @xmath66 it also holds : @xmath67 @xmath68 using together and , we get a system of @xmath69 linear algebraic equations @xmath70 using the boundary conditions , we find @xmath71 from the initial condition we get @xmath72 let @xmath73 @xmath74 and @xmath75 substituting in , we obtain the following system of equations : for @xmath45 , @xmath76 for @xmath77 , @xmath78 for @xmath49 , @xmath79 where @xmath80 @xmath81 @xmath82 in this section we give an example of a model of the thermistor problem : @xmath83 @xmath84 the exact solution of the electrical potencial problem is @xmath85 , @xmath86 . then , the diffusion equation can be reduced to the form @xmath87 using the proposed galerkin finite element approach , we get the following system of algebraic equations : for @xmath45 , @xmath88 for @xmath89 , @xmath90 for @xmath49 , @xmath91 where @xmath92 we now show some results from numerical experiments performed using our method and the computer algebra system . according with physical situations , we choose values of @xmath10 and @xmath93 verifying @xmath94 . in particular , we fixed @xmath95 and @xmath96 . the calculation of the steady - state for the thermistor problem is an important issue regarding the applications of the model in the thermistor device . we obtained stable steady - state times for @xmath97 ( see fig . [ fig : graph ] ) . the authors are grateful to the support of the _ portuguese foundation for science and technology _ ( fct ) through the _ centre for research in optimization and control _ ( ceoc ) of the university of aveiro , cofinanced by the european community fund feder / poci 2010 , and the project sfrh / bpd/20934/2004 . e. caglioti , p .- lions , c. marchioro , m. pulvirenti , a special class of stationary flows for two - dimensional euler equations : a statistical mechanics description , comm . * 143 * ( 1992 ) , no . 3 , 501525 . a. el hachimi and m. r. sidi ammi , existence of weak solutions for the thermistor problem with degeneracy , in _ proceedings of the 2002 fez conference on partial differential equations _ , 127137 ( electronic ) , electron . j. differ . conf . , 9 , southwest texas state univ . , san marcos , tx , 2002 .	_ we use a finite element approach based on galerkin method to obtain approximate steady state solutions of the thermistor problem with temperature dependent electrical conductivity . _ * keywords : * parabolic equation , finite element method , thermistor problem . * 2000 mathematics subject classification : * 35k40 , 74s05 .
phenomena involving positrons are important in many fields of physics , including astrophysics , plasma physics , atomic physics and materials science . in the laboratory , low energy positrons are now being used for many of these applications , including study of electron - positron plasma phenomena @xcite , atomic and molecular physics @xcite , antihydrogen formation @xcite , modeling of astrophysical processes @xcite , and the characterization of materials @xcite . the limitations in these studies are often due to the relative inavailability of suitable positron sources . methods to accumulate , cool and manipulate positron plasmas have been developed @xcite and used in a recent experiment that created low energy antihydrogen atoms at cern @xcite . long confinement times ( days ) , increased plasma densities and brightness enhancement are achieved with the rotating electric field technique @xcite at temperatures of order 1 mev ( @xmath0 10 k ) . the apparatus presented in this article is designed to produce fluxes of slow positrons of order @xmath1 per second and be stored and/or cooled in a trap . in contrast , radioisotope sources currently in use produce fluxes @xmath2 . its size is also much smaller than that of a large linac or of a nuclear reactor but could be used as a small facility for interdisciplinary experimental studies with positrons . such a source would allow the production of positronium ( ps ) , hydrogen and antihydrogen in a symmetric way in the same experimental conditions through the set of reactions : @xmath3 , followed by @xmath4 or its antimatter counterpart @xmath5 and the production of ions via @xmath6 or @xmath7 @xcite . a novel method to produce a 3d image of molecules with a resolution of few ngstrms has been proposed @xcite . since several years the possibility to create a bose - einstein condensate ( bec ) of positronium and to make a 511 kev gamma ray laser is being studied @xcite @xcite . these techniques relie or would benefit a lot on the availability of a very intense source of positrons . moreover a high production rate of slow positrons ( exceeding @xmath8 ) and of positronium is being looked for in industrial and research applications , for instance , `` positron annihilation spectroscopy '' ( pas ) @xcite . the most commonly used source of positrons is @xmath9 . such compact sources are well suited for laboratory research , but their maximum activity lies around @xmath10 bq with a mean lifetime of 2.6 years . there are also some accelerators ( 100 mev ) partly used for the production of slow positrons which are managed as a facility and a nuclear research reactor in munich @xcite . we propose an alternative intense ( @xmath11 ) source of slow ( mev ) positrons based on @xmath12 pair creation through the interaction of a 10 mev electron beam on a target with an intensity of a few ma . this source was designed to be coupled with a greaves - surko trap @xcite in order to produce a bright beam of slow positrons ( mev to kev ) . it may also be used to produce positronium by applying the beam onto a cristal @xcite . the pair production cross section increases with energy . however the development of such a setup for university or industrial applications limits the beam energy to 10 mev because a higher energy would start to activate the environment ( legal limit ) . a dedicated facility would have a size comparable to that of a radioactive source but requires a special building for shielding . the positrons are produced by the interaction of a flat electron beam with a 50 microns target foil . the angle between the beam plane and the foil is very small , approximatively 3 degrees . the positron kinetic energy spectrum is peaked at 1.2 mev and extends to 8 mev . the first step in the positron capture by the trap is the moderation process . this process involves the slowing down of the positrons , the creation of meta - stable states with collective charge oscillations in the moderator and its re - emission at @xmath0 1.5 ev . the moderation efficiency decreases with the incident positron kinetic energy and is negligible at a few mev . therefore a magnetic collector was designed to separate the positrons from the electrons while preserving the positrons with a kinetic energy below 1 mev . an experimental test was performed with an electron gun used to weld metal pieces . the gun delivered a beam with a diameter around 5 mm , i.e. a surface of around 20 @xmath13 ( fig [ fig : soudure1 ] ) . the 99.99% purity tungsten target we tested had a thickness of 50 @xmath14 m @xcite , dimensions of 5 cm x 5 cm and was held with a piece of tungsten surrounding it on three of its sides . the accelerating voltage of the gun was fixed at 40 kv , the intensity was gradually increased until perforation at 20 ma . a 15 ma current does not perforate the target . at 40 kv electrons deposit all their energy in the metal . the target sustains thus a deposit greater than 2 kw/@xmath15 . however a 1 kw/@xmath15 limit is set on the deposited power as a safety margin . even if the temperature is kept well below the fusion point , there will be metal evaporation . using tables from langmuir and jones @xcite , on the evaporation rate of tungsten filaments for light bulbs under the joule effect , the target would lose 10% of its mass in one hour at 3100 k , and 24 hours at 2700 k. a simple way of operation would then consist in exchanging the tungsten foil every night . running at 2700 k means a lower electron intensity and thus a reduction in the @xmath16 rate of a factor 1.7 . simulations were performed with the geant 3.21 @xcite fortran library . electrons of 10 mev / c momentum were generated at an incidence angle of 3 or 90 degrees w.r.t the target plane . for the 3 degree incidence angle , the transverse shape of the beam is a rectangle of 1 mm x 20 mm thus illuminating a square area of 20 mm x 20 mm on the target . we have computed the energy deposited in targets of various thicknesses with a current of 1 ma ( figure [ fig : courantmax ] ) . for 50 @xmath14 m at 3 degrees or 1 mm at 90 degrees , which have the same equivalent thickness ) . ] , the deposited energy is respectively 1750 w and 4500 w. this difference is due to the fact that electrons have more possibilities to escape the target when the incidence angle is small . the total path length of an electron and its secondary electrons inside the target is on average 5 times longer at 90 degrees than at 3 degrees for the same equivalent thickness . the deposited energy increases with equivalent thickness but saturates earlier with very low incidence angle . this allows to deliver a higher intensity for the same illuminated surface . figure [ fig : courantmax ] shows the electron current intensity which corresponds to a deposited energy of 1 kw as a function of the equivalent thickness . the positron rate is given in figure [ fig : prod ] . it is of the order of @xmath17 for 1ma electron current and equivalent thicknesses between 1 and 2 mm . these results agree with an independant similar study @xcite . the rate of positrons produced with less than 1 mev of kinetic energy shown in figure [ fig : prod ] is about 1/5 of the total rate . the limit on the deposited power at 1 kw/@xmath15 , determines the maximum current intensity per @xmath15 of target ( see figure [ fig : courantmax ] ) . the corresponding positron rates are shown in figure [ fig : ratepos-1kw ] . let us take the example of a 1 @xmath15 target of 50 micron thickness at an incidence of 3 degrees or 0.96 mm equivalent thickness . the deposited power for a 1 ma electron current is 1.75 kw ( resp . 4.5 kw ) at 3 degrees ( resp . 90 degrees ) . the maximum acceptable current for the limit of 1 kw/@xmath15 , as well as the corresponding positron rate for the maximum current are given in table [ tab : tab1 ] . a system of production and collection of mev positrons has been presented . this setup uses pair creation from electrons hitting a thin tungsten target at a 3 degree incidence angle . the beam comes from a 10 mev/2.3 ma electron accelerator running in continuous mode . the setup allows to collect more than 5 @xmath19 positrons of less than 600 kev in a 2 cm radius aperture at 2 m from the target . the flux of electrons and photons reaching this aperture is of the order of 50 w. the best electron source for this application is a rhodotron , which is commercialized by the iba @xcite firm . this compact machine ( about 3 m in diameter ) is operated continuously all year long and can reach beam intensities up to 100 ma for food disinfection . several models are available according to the desired beam intensity . a 10 mev/2ma model would use 250 kw of power . * acknowledgements * + we wish to express our sincere thanks to all the people from different fields with whom we had fruitful discussions in order to design this project : b. aune , g. baldacchino , j.l . borne , p. debu , s. ecklund , r. greaves , m. jablonka , p. lecoeur , b. mansouli , f. mot , j.c . mialocq , a.p . mills , a. nguyen , j. sheppard , m. spiro , c. surko , g. vigneron , and m. woods . murphy and c.m . surko , phys . rev . * a 46 * ( 1992 ) 5696 ; c.m . surko , s.j . gilbert and r.g . greaves , non - neutral plasma phys . iii , edited by j.j . bollinger , r.l . spencer and r.c . davidson , ( american institute of physics , new york ) , 3 - 12 ( 1999 ) ; + http://physics.ucsd.edu/research/surkogroup/positron/buffergas.html m. charlton , and j.w . humbertson , positron physics , ( cambridge univ . press , 2001 ) ; c.m . surko , new directions in antimatter chemistry and physics , surko and gianturco eds , 2001 , kluwer academic publishers . m. amoretti et al . , nature , * 419 * ( 2002 ) 456 ; g. gabrielse et al . , * b 548 * ( 2002 ) 140 ; g. gabrielse et al . , phys . lett . * 89 * , 233401 ( 2002 ) ; g. gabrielse et al . * 89 * , 213401 ( 2002 ) .	we describe a novel design for an intense source of slow positrons based on pair production with a beam of electrons from a 10 mev accelerator hitting a thin target at a low incidence angle . the positrons are collected with a set of coils adapted to the large production angle . the collection system is designed to inject the positrons in a greaves - surko trap @xcite . such a source could be the basis for a series of experiments in fundamental and applied research and would also be a prototype source for industrial applications which concern the field of defect characterization in the nanometer scale . positron , positronium , accelerator / linac , defect characterization , 3d molecule imaging , energy storage , gamma ray laser , matter antimatter symmetry , antigravity code , code
in the introductory paragraph of their recent interesting letter , sjqvist _ et al _ state that `` uhlmann was probably the first to address the issue of mixed state holonomy , but as a _ purely mathematical _ ( emphasis added ) problem . in contrast , here we provide a new formalism of the geometric phase for mixed states in the _ experimental _ ( emphasis added ) context of quantum interferometry '' @xcite . left unaddressed there , however , was the nature of any possible relationship between the two approaches . this question merits particular attention since dittmann has `` shown that the connection form ( gauge field ) related to the generalization of the berry phase to mixed states proposed by uhlmann satisfies the source - free yang - mills equation @xmath0 , where the hodge operator is taken with respect to the bures metric on the space of finite - dimensional nondegenerate density matrices '' @xcite . this fact , dittmann and uhlmann noted , `` may be seen as an extension to mixed states of numerous examples relating the original berry phase to dirac monopoles , and the wilczek and zee phase , to instantons '' @xcite . in their work , et al _ do express the geometric phase in terms of an _ average _ connection form ( * ? ? ? ( 16 ) ) ( for which no yang - mills characterization has been given ) . zee has reported that the non - abelian gauge structure associated with nuclear quadrupole resonance does _ not _ satisfy the source - free yang - mills equation @xcite ( cf . @xcite ) . in @xcite , in line with certain work of figueroa - ofarrill concerning calibrated geometries @xcite , we have begun a study of the `` dittmann - bures / yang - mills '' field over the _ eight_-dimensional convex set of _ three_-level density matrices ( making use of the formulas for the bures metric developed in @xcite , based on the parameterization in @xcite ) by finding certain self - dual _ four_-forms ( @xmath1 ) with respect to the bures metric . the eigenspace decomposition of the associated endomorphism @xmath2 was found to yield four octets and three singlets as factors @xcite . the research presented below represents our effort to gain a formal understanding of the relations between the `` mathematical '' methodology of uhlmann and the `` experimental '' approach of sjqvist _ et al_. we hope that this will prove useful , among other things , in the devising of a physical scheme for implementing the concepts of uhlmann ( cf . @xcite ) . in the interferometric / experimental case @xcite , each pure state that diagonalizes the initial density matrix ( @xmath3 ) is parallel transported separately from other distinct diagonalizing pure states , so one simply weights or ( incoherently ) averages the individual pure state ( single - state ) interference profiles by the corresponding eigenvalues of @xmath3 to obtain the mixed state results . in the notation of @xcite , the _ total _ geometric phase is the argument of @xmath4 , and the visibility is its absolute value , where @xmath5 are , respectively , the eigenvalue , visibility and geometric phase obtained for the @xmath6-th constituent pure state of @xmath3 ( * ? ? ? ( 10 ) , ( 15 ) ) . the phase introduced by sjqvist _ et al _ `` fulfills two central properties that make it a natural generalization of the pure state case : ( i ) it gives rise to a linear shift of the interference oscillations produced by a variable @xmath7 phase and ( ii ) it reduces to the pancharatnam connection for pure states '' @xcite . the mixed state generalization of pancharatnam s connection which asserts that @xmath8 and @xmath9 are in phase if @xmath10 $ ] is real and positive can be met only when ( * ? ? ? * eq . ( 12 ) ) @xmath11=0.\ ] ] this is the parallel transport condition of sjqvist _ et al _ for mixed states undergoing unitary evolution . in the approach of uhlmann @xcite the parallel transport takes place in a larger hilbert space , defined by purification . the geometric phase of sjqvist _ et al _ can also be understood using purification @xcite , though the unitary operator that acts on the ancilla or auxiliary state there is unconstrained and can in fact be anything , such as the identity operator . a parallel transport of a density operator amounts to a parallel transport of _ any _ ( emphasis added ) of it purifications @xcite . in infinitesimal form , the parallelity condition of uhlmann for a density matrix @xmath8 reads ( * ? ? ? * eq . ( 19 ) ) , @xmath12 where @xmath13 . the unitary operator @xmath14 can be interpreted as acting on the ancilla in the purification of @xmath8 and is determined by this parallelity condition ( [ w1 ] ) . for example , if the system , initially in the state @xmath3 , evolves unitarily under the evolution operator @xmath15 one may write @xmath16 . the uhlmann parallelity condition , then , reads @xmath17 with given input @xmath3 and @xmath15 this equation needs to be solved for @xmath14 . the geometric phase @xmath18 for any specific @xmath19 can then be computed using the mixed state pancharatnam connection @xcite , @xmath20 in the three sets of analyses reported below ( sec . ii ) , we will implement certain formulas of uhlmann @xcite . we compare the two methodologies under consideration in the context of : ( a ) unitary evolution over geodesic spherical _ triangles _ , the vertices of which are associated with certain specific _ mixed _ two - level systems ; ( b ) unitary evolution of such two - level systems over certain circular paths ( @xmath21-orbits ) ; and ( c ) unitary evolution over such circular paths of @xmath6-level gibbsian density matrices @xcite . in case ( a ) we find an interesting correspondence between the two approaches for mixed states having bloch vectors of length @xmath22 and in ( c ) that the results of the uhlmann methodology converge to those of sjqvist _ et al _ if a certain variable ( a ) , a function ( @xmath23 ) of the inverse temperature parameter ( @xmath24 ) , is driven to zero . if instead of strictly mixed states , we consider the unitary evolution of _ pure _ states , then both methodologies simply reproduce the corresponding berry phase for each of the three scenarios . however , for general mixed states , the two methodologies typically yield inequivalent results . for a qubit ( a spin-@xmath25 particle ) , whose density matrix can be written as , @xmath26 where @xmath27 is a unit vector , @xmath28 is the constant ( length of the bloch vector ) for unitary evolution and @xmath29 is a vector of the three ( non - identity ) pauli matrices , sjqvist _ et al _ obtained a formula for the _ geometric phase _ * eq . ( 25 ) ) , @xmath30 = - \mbox{arctan } \big ( r \mbox{tan } { \omega \over 2 } \big).\ ] ] here @xmath31 is the unitary curve in parameter space traversed by the system and @xmath32 the solid angle subtended by the natural extension of @xmath31 to the unit ( bloch ) sphere . this formula reduces to @xmath33 for pure states ( @xmath34 ) , as it , of course , should . the _ visibility _ was expressed as ( * ? ? ? * eq . ( 26 ) ) , @xmath35 for the case of cyclic evolution , which will be our only area of concern here , @xmath36 . we focus on the case of cyclic evolution for geodesic triangles , in particular , since uhlmann has an enabling formula `` relevant for the intensities '' , ( * ? ? ? * eq . ( 24 ) ) @xmath37 @xmath38 here @xmath39 where @xmath40 gives the projection operator @xmath41 associated with the @xmath42-th density matrix ( @xmath43 ) in the geodesic triangle , that is , @xmath44 uhlmann also apparently has an explicit formula for geodesic _ quadrangles _ , but only the leading terms of it have been published ( * ? ? ? ( 25 ) ) , that is , @xmath45 @xmath46 without loss of generality , let us set one of the three vertices of the geodesic spherical triangle swept out by the unit vector @xmath27 ( [ fee ] ) in the unitary evolution to be ( 0,0,1 ) . we use the standard spherical coordinate form for a point on the sphere , that is @xmath47 , and take the other two vertices to be parameterized by the angular pairs @xmath48 and @xmath49 , so the preassigned vertex ( 0,0,1 ) corresponds simply to @xmath50 . then , using the notation @xmath51 , we have for the solid angle @xcite , @xmath52 ( this formula , as well as all the succeeding ones here , depend implicitly or explicitly upon @xmath53 and not on @xmath54 and @xmath55 individually . ) let us also denote @xmath56 the parameter @xmath57 is confined to the range @xmath58 $ ] , for @xmath59 . it is equal to its minimum -@xmath25 at @xmath60 and @xmath61 ; equal to its maximum 4 for @xmath62 and equal to zero if _ either _ @xmath63 or @xmath64 . making use of ( [ armin1 ] ) and ( [ armin2 ] ) , we have for the `` uhlmann geometric phase '' for the general geodesic triangle specified that @xmath65 where @xmath66 we found that @xmath67 so one can write ( [ one1 ] ) @xmath68 as well as @xmath69 let us note the _ exact _ relations between the two forms ( `` mathematical '' and `` experimental '' ) of geometric phase , @xmath70 which are unity for both @xmath71 and @xmath72 . ( anandan _ et al _ @xcite assert that `` a fully gauge - invariant description of the geometric phase _ requires _ that it be defined modulo @xmath73 '' . of course , @xmath74 . ) for @xmath75 , @xmath76 for @xmath77 , @xmath78 for @xmath79 ( that is , a neighborhood of the fully mixed [ classical ] state , @xmath80 ) , @xmath81 the `` uhlmann visibility '' is @xmath82 which is precisely unity for @xmath71 , as is also the case for @xmath83 , given by ( [ sjovis ] ) . now , @xmath84 this ratio is always unity for @xmath71 and always @xmath85 for @xmath86 . ( recall , on the other hand , that the ratio ( [ ratio ] ) of the tangents of the two geometric phases is _ unity _ for @xmath87 . ) for @xmath88 , @xmath89 for @xmath90 , @xmath91 for @xmath92 , @xmath93 in the limit @xmath94 , @xmath95 now let us similarly study the unitary evolution of spin-@xmath25 systems over _ circular _ ( @xmath21-orbits ) as opposed to geodesic triangular paths . we utilize the analyses of uhlmann in @xcite . again , we start with an initial density matrix @xmath3 , the bloch vector of which is proportional to ( 0,0,1 ) , while the axis of rotation about which the state evolves is @xmath96 . the resulting curve of density matrices is given by @xmath97 and the associated parallel lift of this curve with initial value @xmath98 by @xmath99 where @xmath100 and the @xmath101 s are , of course , generators of an irreducible representation of @xmath102 . we have the relation @xmath103 the associated holonomy invariant for a complete rotation about @xmath104 is @xmath105 ( in @xcite , we have studied certain `` higher - order '' holonomy invariants . ) we then have the result ( using the `` @xmath106 '' notation now , rather than `` @xmath107 '' as before , to refer to results obtained based on the unitary circular evolution analysis of uhlmann @xcite ) @xmath108 where @xmath109 and @xmath110 we plot this ratio ( [ par1 ] ) in fig . [ lb ] . now for @xmath111 , @xmath112 for pure states ( @xmath71 ) , the ratio ( [ par1 ] ) reduces to 1 . the ratio of the visibilities , @xmath113 is given in fig . [ lc ] . both the ratios ( [ par1 ] ) and ( [ par2 ] ) are unity for pure states ( @xmath71 ) . in the final part of the paper @xcite , uhlmann also considered the analogous unitary circular evolution ( @xmath21-orbits ) of @xmath6-level gibbsian states of the form @xmath114 where the vector @xmath115 of angular momentum operators lives in @xmath6-dimensions . ( for @xmath116 , the model is equivalent to that studied in the previous section . ) in @xcite we have conducted a detailed analysis of this model in comparison with the results that would be obtained implementing the methodology of sjqvist _ et al_. one unanticipated outcome was that the sjqvist _ et al _ model appeared sensitive to the bosonic ( odd @xmath6 ) or fermionic ( even @xmath6 ) nature of the @xmath6-level systems . as an example of this phenomenon , we exhibit two additional plots ( fig . [ p1 ] and [ p2 ] ) based on analyses for @xmath117 . in each of these plots , the inverse temperature parameter @xmath24 is held fixed at 2 . in fig . [ p1 ] , based on the uhlmann methodology , below @xmath118 the ten curves are all monotonically decreasing . the order of dominance among themselves of these curves is the natural order @xmath117 , with the curve for @xmath116 being the dominant one . on the other hand , using the procedure of sjqvist _ et al _ , the ten corresponding curves are separated into two clusters of five ( below @xmath119 ) ( fig . [ p1 ] ) . in the first cluster ( having the higher values of @xmath120 ) in decreasing order of dominance the five curves correspond to @xmath121 , while in the second group ( having lesser values ) the order is @xmath122 . ( for other related figures , see @xcite . ) [ y3 ] also , we have found @xcite that the ( first - order ) holonomy invariant in the uhlmann approach ( * ? ? ? ( 33 ) , ( 43 ) , ( 47 ) ) , that is the trace of @xmath123 reduces to @xmath124 times the holonomy invariant ( * ? ? ? * eq . ( 15 ) ) in the sjqvist _ et al _ analysis if we simply set the parameter @xmath125 in ( [ q1 ] ) to zero , that is the influence of the angular momentum operator @xmath126 is nullified . relatedly , if @xmath3 were the density matrix of a _ pure _ state , then the term @xmath127 in the invariant ( [ q1 ] ) would drop out , with the trace of the resultant expression being the corresponding berry phase ( * ? ? * eq . ( 43 ) ) . this fact was employed in @xcite to compute the sjqvist _ et al _ mixed state holonomy for the @xmath6-level gibbsian density matrices , by taking the average ( using the eigenvalues of @xmath3 as weights ) of these berry phases for the pure states that diagonalize the initial density matrix ( * ? ? ? ( 10 ) , ( 15 ) ) . the analyses presented above clearly demonstrate a general inequivalence between the `` experimental '' approach to mixed state holonomy of sjqvist _ et al _ @xcite and the `` purely mathematical '' scheme of uhlmann @xcite . in particular , in the third of our analyses ( sec . [ ig ] ) , concerned with @xmath21-orbits of @xmath6-level gibbsian density matrices , it has emerged that the analyses of sjqvist _ et al _ ( fig . [ p2 ] ) are sensitive to the parity of @xmath6 , while those of uhlmann ( fig . [ p1 ] ) are not @xcite . ( for an application of @xcite to the detection of quantum entanglement , see @xcite . for a further discussion of the results in @xcite , see @xcite . ) a problem that remains to be fully addressed is the conceptualization of a physical apparatus that would yield the geometric phases and visibilities given by the analyses of uhlmann @xcite . in their study @xcite , sjqvist _ et al _ did describe the use of a mach - zehnder interferometer to `` test the geometric phase for mixed states and provide the notion of phase and parallel transport for mixed states in a straightforward way '' . however , they did not specify such a physical scheme for their conceptual alternative based on a purification procedure ( cf . @xcite ) . let us also note that the connection used for parallel transport in the uhlmann ( bures ) case is but one of an ( uncountable ) infinite set of connections ( `` defining reasonable parallel transports along curves of density operators '' ) , which correspond in a one - to - one fashion , with the _ monotone _ metrics on the @xmath6-level density matrices , and are all compatible with the purification procedure @xcite . in this generalized framework , the uhlmann ( canonical ) connection corresponds to the well - studied bures or minimal monotone metric @xcite . a yet apparently formally unsettled question is whether or not the sjqvist _ et al _ connection ( * ? ? ? ( 16)-(18 ) ) itself corresponds to some such monotone metric , thus falling within this generalized scheme of dittmann and uhlmann of connections ( and metrics ) respecting standard purification @xcite . from the results presented above , it is clear that it does _ not _ correspond to the bures metric . ( in the uhlmann methodology , `` proper liftings '' @xmath128 are those that minimize the length functional @xmath129 $ ] of the curve of density operators undergoing unitary evolution . for a proper lifting this length becomes equal to the `` bures length '' . one particularly interesting property of the bures metric is that its volume element can be employed to determine analytically _ exact _ probabilities that a pair of quantum bits is [ _ a priori _ ] classically correlated @xcite . ) sjqvist e. , pati , a. k. , ekert , a. , anandan , j. s. , ericsson , m. , oi , d. k. l. , and vedral , v. : geometric phases for mixed states in interferometry . _ 85 , 2845 ( 2000 ) . dittmann , j. : yang - mills equation and bures metric . 46 , 281 ( 1998 ) . dittmann , j. and uhlmann , a. : connections and metrics respecting purification of quantum states . j. math . 40 , 3246 ( 1999 ) . zee , a. : non - abelian structure in nuclear quadrupole resonance . a , 38 , 1 ( 1988 ) . avron , j. e. , sadun , l. , segert , j. and simon , b. : topological invariants in fermi systems with time - reversal invariance . 61 , 1329 ( 1988 ) . slater , p. b. : self - duality , four - forms , and the eight - dimensional yang - mills field over the three - level density matrices . e - print , math - ph/0108005 . figueroa - ofarrill , j. m. : intersecting brane geometries . 35 , 99 ( 2000 ) . slater , p. b. : bures geometry of the three - level quantum systems , j. geom . 39 , 207 ( 2001 ) . byrd , m. and slater , p. b. : bures measures over the spaces of two and three - dimensional density matrices . phys . lett . a 283 , 152 ( 2001 ) . sjqvist , e.:pancharatnam revisted . e - print , quant - ph/0202078 . uhlmann , a. : on cyclic evolution of mixed states in two - level systems . in _ nonlinear , dissipative , irreversible quantum systems _ ( clausthal , 1994 ) , edited by doebner , h .- d . , dobrev , v. k. , and natterman , p. ( world scientific , 1995 ) , p. 296 . uhlmann , a. : parallel lifts and holonomy along density operators : computable examples using @xmath21-orbits . in _ symmetries in science vi _ , edited by b. gruber , ( plenum , new york , 1993 ) , p. 741 . slater , p. b. : geometric phases of the uhlmann and sjqvist _ et al _ types for @xmath21-orbits of @xmath130-level gibbsian density matrices . e - print , math - ph/0112054 . eriksson , f. : on the measure of solid angles . math . mag . 63 , 184 ( 1990 ) . anandan , j. , sjqvist , e. , pati , a. k. , ekert , a. , ericsson , m. , oi , d. k. l. , and vedral , v. : reply to `` singularities of the mixed state phase '' . e - print , quant - ph/0109139 . horodecki , p. and ekert , a. : direct detection of quantum entanglement . e - print , quant - ph/0111064 . petz , d. and sdar , c. : geometries of quantum states . j. math . 37 , 2662 ( 1996 ) . slater , p. b. : monotonicity properties of certain measures over the two - level quantum systems . 52 , 343 ( 2000 ) . dittmann , j. : explicit formulae for the bures metric . j. phys . a 32 , 2663 ( 1999 ) . slater , p. b. : exact bures probabilities that two quantum bits are classically correlated . j.b 17 , 471 ( 2000 ) .	sjqvist , pati , ekert , anandan , ericsson , oi and vedral ( phys . rev . lett . 85 , 2845 [ 2000 ] ) have recently `` provided a physical prescription based on interferometry for introducing the total phase of a mixed state undergoing unitary evolution , which has been an elusive concept in the past '' . they note that `` uhlmann was probably the first to address the issue of mixed state holonomy , but as a purely mathematical problem '' . we investigate possible relationships between these `` experimental '' and `` mathematical '' approaches , by examining various quantum - theoretic scenarios . we find that the two methodologies , in general , yield inequivalent outcomes . -0.1 cm * mathematics subject classifications ( 2000 ) * : 81q70 * key words . * geometric phase , berry phase , mixed states , density matrix , bures metric , geodesic triangles , gibbsian density matrices , yang - mills connection , uhlmann phase
hohenberg and kohn ( hk ) replaced the wave function by the electronic density @xmath11 as the fundamental variable of the many - body problem and thereby achieved a crucial breakthrough in the theoretical description of the electronic properties of matter.@xcite since then , density functional theory ( dft ) has been the subject of remarkable developments . formal improvements , extensions , and uncountable successful applications to a large variety of problems have made of this theory the most efficient , albeit not infallible , method of determining physical and chemical properties of matter from first principles.@xcite density functional ( df ) calculations are usually based on the kohn - sham ( ks ) scheme which reduces the correlated @xmath12-particle problem to the solution of a set of self - consistent single - particle equations.@xcite while this transformation is formally exact , the form of the interaction - energy functional @xmath13 $ ] involved in the ks equations is not known explicitly . practical implementations of dft always require approximations to @xmath13 $ ] , or equivalently to the exchange and correlation ( xc ) functional @xmath14 $ ] , on which the quality of the results depends crucially . therefore , understanding the functional dependence of @xmath13 $ ] and exploring new ways of improving its approximations are central to the development of df methods . the most extensively used forms for @xmath13 $ ] are presently the local density approximation ( lda ) , @xcite its spin - polarized version or local spin - density approximation ( lsda)@xcite and the gradient corrected extensions,@xcite which were originally derived from exact results for the homogeneous electron gas . despite an unparalleled success in the most diverse areas , the lda - based approach fails systematically in accounting for phenomena that involve strong electron - correlation effects as observed , for example , in mott insulators , heavy - fermion materials , or high-@xmath15 superconductors . these systems are usually described in the framework of parametrized lattice hamiltonians such as anderson,@xcite hubbard,@xcite pariser - parr - pople,@xcite and related models which focus on the most relevant electron dynamics at low energies . however , even with simplified model interactions and a minimal number of orbitals per atom , a detailed understanding of the electronic properties in the strongly correlated limit remains a serious theoretical challenge . exact results are rare or numerically very demanding and a variety elaborate many - body techniques have been specifically developed in order to study this problem.@xcite being in principle an exact theory , the limitations of the df approach have to be ascribed to the approximations used for exchange and correlation and not to the underlying formalism . it is therefore very interesting to extend the range of applicability of dft to the many - body lattice models that describe the physics of strongly correlated systems . moreover , the development of lattice density - functional theory ( ldft ) constitutes an intrinsically inhomogeneous approach and provides a true alternative to the lsda and related gradient - corrected methods . thus , studies on simple models can open new insights into the properties of @xmath6 that should also be useful for future extensions to more realistic hamiltonians and first principles calculations . several physical problems have been already investigated by applying the concepts of dft to lattice models , for example , the band - gap problem in semiconductors,@xcite the role of off - diagonal elements of the density matrix and the non - interacting @xmath16 representability in strongly correlated systems,@xcite or the development of energy functionals of the density matrix with applications to hubbard and anderson models.@xcite in previous works@xcite we have considered a density - matrix functional theory of many - body lattice models , that is analogous to gilbert s approach in the continuum,@xcite and applied it to the hubbard hamiltonian with uniform nearest - neighbor ( nn ) hopping integrals @xmath17 . the interaction energy @xmath6 of the hubbard model has been calculated exactly as a function of the density matrix @xmath0 for various periodic lattices having @xmath18 for all nearest neighbors @xmath4 and @xmath19 . an analysis of the functional dependence of @xmath20 for different band fillings and lattice structures revealed very interesting scaling properties.@xcite on this basis , a simple general approximation to @xmath20 has been derived which yields a remarkable agreement with available exact results in 1d systems and which predicts successfully the ground - state energy and charge - excitation gap of the 2d hubbard model in the complete range of interaction strength.@xcite this shows that dft with an appropriate approximation to @xmath6 is an efficient tool for determining the electronic properties of many - body lattice models . the purpose of this paper is to extend the method by allowing for alternations of the density - matrix elements @xmath0 between nearest neighbors in order to study the dimerized 1d hubbard model . this problem has motivated a considerable research activity in past years , particularly concerning the role of electron correlations in the dimerization of polymer chains like polyacetylene.@xcite in this context two qualitatively different regimes may be distinguished depending on the relative importance of the intra - atomic coulomb repulsion @xmath21 and the nn hopping integral @xmath22 . on the one side , for small @xmath7 , the dimerization can be regarded as a bond - order wave that opens a gap at the fermi surface of the one - dimensional ( 1d ) single - particle band - structure ( peierls distortion ) . on the other side , for large @xmath7 , local charge fluctuations are severely reduced and the low - energy properties are dominated by antiferromagnetic ( af ) correlations between spin degrees of freedom . in this case the dimerization can be regarded as an alternation of the strength of af correlations along the chain or spin - peierls state . one of our aims is to analyze the differences between these two types of behaviors in the framework of ldft . the properties of dimerized chains are also very interesting from a purely methodological point of view . they provide in fact a simple , physically motivated means of exploring the functional dependence of @xmath1 $ ] by including additional degrees of freedom , thereby allowing for a larger flexibility . moreover , several exact results are available to compare with ( e.g. , bethe - ansatz solution for the non - dimerized hubbard chain , finite - ring lanczos diagonalizations , or density - matrix renormalization - group calculations ) which allow to quantify the accuracy of the final results . the remainder of the paper is organized as follows . in sec . [ sec : ldft ] the main steps in the formulation of ldft are briefly recalled . the properties of the interaction - energy functional @xmath6 of the dimerized hubbard model are discussed in sec . [ sec : w ] . the domain of representability of @xmath0 and the scaling behavior of @xmath6 are investigated . a simple explicit approximation to @xmath6 is derived , which is appropriate for direct calculations . [ sec : dimhub ] is concerned with applications . the ground - state energy and the charge - excitation gap of finite and infinite dimerized chains are determined as a function of coulomb - repulsion strength @xmath7 and hopping - integral alternation @xmath8 . the ldft results are contrasted with accurate numerical solutions in order to discuss goals and limitations of the present approach . finally , sec . [ sec : concl ] summarizes the main conclusions and points out some future perspectives . we consider the many - body hamiltonian @xmath23 where @xmath24 ( @xmath25 ) is the usual creation ( annihilation ) operator for an electron with spin @xmath26 at site or orbital @xmath4 . the hopping integrals @xmath9 define the lattice ( e.g. , 1d chains , square or triangular 2d lattices ) and the range of single - particle interactions ( e.g. , up to first or second neighbors ) . from the _ ab initio _ perspective @xmath9 is given by the external potential @xmath27 and by the choice of the basis . @xmath28 defines the type of many - body interactions which may be repulsive ( coulomb like ) or attractive ( in order to simulate electronic pairing ) and which are usually approximated as short ranged ( e.g. , intra - atomic ) . ( [ eq : hamgen ] ) is mainly used in this section to present the general formulation which can then be applied to various specific models by simplifying the interactions . a particularly relevant example , to be considered in some detail in following sections , is the single - band hubbard hamiltonian with nn hoppings.@xcite the non - dimerized form of this model is obtained from eq . ( [ eq : hamgen ] ) by setting @xmath29 for @xmath4 and @xmath19 nn s , @xmath30 otherwise , and @xmath31 . the hopping matrix @xmath9 plays the role given in conventional dft to the external potential @xmath32 . consequently , the single - particle density matrix @xmath0 between lattice sites replaces the continuum density @xmath11 as basic variable . the situation is similar to the density - matrix functional theory proposed by gilbert for the study of non - local pseudo - potentials @xmath33.@xcite the ground - state energy @xmath34 and density matrix @xmath35 are determined by minimizing the energy functional @xmath36 = e_k[\gamma_{ij } ] + w [ \gamma_{ij}]\ ] ] with respect to @xmath0 . @xmath37 $ ] is defined for all density matrices that can be written as @xmath38 for all @xmath4 and @xmath19 , where @xmath39 is an @xmath12-particle state . in other words , @xmath0 derives from a physical state and is said to be pure - state @xmath12-representable.@xcite the first term in eq . ( [ eq : efun ] ) is given by @xmath40 it represents the kinetic energy associated with the electronic motion in the lattice and includes all single - particle contributions . notice that eq . ( [ eq : ek ] ) yields the exact kinetic energy and that no corrections on @xmath41 have to be included in other parts of the energy functional as in the ks approach . the second term in eq . ( [ eq : efun ] ) is the interaction - energy functional given by@xcite @xmath42 = min \left[\frac{1}{2 } \sum _ { klmn \atop { \sigma \sigma ' } } v_{klmn } \ ; \langle \psi [ \gamma_{ij } ] \| \ ; \hat c_{k \sigma } ^{\dagger } \hat c_{m \sigma ' } ^{\dagger } \hat c_{n \sigma ' } \hat c_{l \sigma } \ ; \|\psi[\gamma_{ij}]\rangle \right ] \ ; , \ ] ] where the minimization implies a search over all @xmath12-particles states @xmath43 \rangle$ ] that satisfy @xmath44 \| \ ; \sum_\sigma \hat c_{i \sigma } ^{\dagger}\hat c_{j \sigma } \ ; @xmath19 . @xmath1 $ ] represents the minimum value of the interaction energy compatible with a given density matrix @xmath0 . it is often expressed in terms of the hartree - fock energy @xmath45=\frac{1}{2}\sum_{ijkl \atop{\sigma \sigma ' } } v_{ijkl } \left(\gamma_{ij \sigma } \gamma_{kl \sigma ' } - \delta_{\sigma\sigma ' } \gamma_{il \sigma } \gamma_{kj \sigma } \right)\ ] ] and the correlation energy @xmath46 $ ] as @xmath47 = e_{\rm hf}[\gamma_{ij } ] + e_{\rm c}[\gamma_{ij } ] \;.\ ] ] @xmath6 and @xmath48 are universal functionals of @xmath0 in the sense that they are independent of @xmath9 , i.e. , of the system under study . they depend on the considered interactions or model , as defined by @xmath49 , on the number of electrons @xmath50 , and on the structure of the many - body hilbert space , as given by @xmath50 and the number of orbitals or sites @xmath51 . @xmath52 $ ] is minimized by expressing @xmath53 in terms of the eigenvalues @xmath54 ( occupation numbers ) and eigenvectors @xmath55 ( natural orbitals ) as @xmath56 lagrange multipliers @xmath57 and @xmath58 ( @xmath59 ) are introduced in order to impose the usual constraints @xmath60 and @xmath61 . derivation with respect to @xmath62 and @xmath54 ( @xmath63 ) yields the eigenvalue equations@xcite @xmath64 with the following conditions relating @xmath54 and @xmath65 : @xmath66 and @xmath67 self - consistency is implied by the dependence of @xmath68 on @xmath54 and @xmath55 . ( [ eq : minsc])([eq : minepsc ] ) hold exactly in all interaction regimes . they are analogous to well - known results of density - matrix functional theory in the continuum.@xcite however , notice the difference with the ks - like approach considered in ref . @xcite , which assumes non - interacting @xmath16-representability , and where only integer occupations are allowed . the importance of fractional natural - orbital occupations has already been stressed in previous density - matrix functional studies in the continuum.@xcite in fact , in the case of models one observes that @xmath69 for all @xmath70 except in very special situations like the uncorrelated limit ( @xmath71 ) or the fully - polarized ferromagnetic state in the hubbard model . this can be understood from perturbation - theory arguments none of the @xmath54 should be a good quantum number for @xmath72 and has been explicitly demonstrated in exact solutions for finite systems or for the 1d hubbard chain.@xcite therefore , the case ( [ eq : minepsb ] ) is the only relevant one in general . all @xmath65 in eq . ( [ eq : minsc ] ) must be degenerate or equivalently @xmath73 clearly , approximations of @xmath6 in terms of the diagonal @xmath74 alone can never yield such a behavior . given a self - consistent scheme that implements the variational principle , the challenge remains to find good approximations to @xmath1 $ ] that are simple enough to be applied in practical calculations . in order to determine @xmath1 $ ] from eq . ( [ eq : w ] ) we look for the extremes of @xmath75 \ ; + \;\varepsilon \ ; \big(1 - \langle \psi \|\psi \rangle \big ) \ ; \nonumber \\ & + & \sum_{i , j } \lambda_{ij } \ ; \big ( \langle\psi\| \sum_\sigma \hat c^\dagger_{i\sigma } \hat c_{j \sigma } \| \psi\rangle - \gamma_{ij } \big ) \end{aligned}\ ] ] with respect to @xmath76 . lagrange multipliers @xmath77 and @xmath78 have been introduced to enforce the normalization of @xmath79 and the representability of @xmath0 as required by eq . ( [ eq : rep ] ) . derivation with respect to @xmath80 , @xmath77 and @xmath78 yields the eigenvalue equations@xcite @xmath81 and the auxiliary conditions @xmath82 and @xmath83 . the lagrange multipliers @xmath78 play the role of hopping integrals to be chosen in order that @xmath79 yields the given @xmath0 . the pure - state representability of @xmath0 ensures that there is always a solution . the subset of @xmath0 that can be represented by a ground - state of eq . ( [ eq : evgen ] ) for some @xmath78 is the physically relevant one , since it necessarily includes the absolute minimum @xmath35 of @xmath37 $ ] . nevertheless , it should be noted that pure - state representable @xmath0 may be considered that can only be described by excited states or by linear combinations of eigenstates of eq . ( [ eq : evgen]).@xcite the general functional @xmath1 $ ] , valid for all lattice structures and for all types of hybridizations , can be simplified at the expense of universality if the hopping integrals are short ranged . for example , if only nn hoppings are considered , the kinetic energy @xmath41 is independent of the density - matrix elements between sites that are not nn s . therefore , the constrained search in eq . ( [ eq : w ] ) may be restricted to the @xmath43 \rangle$ ] that satisfy eq . ( [ eq : rep ] ) only for @xmath84 and for nn @xmath85 . this reduces significantly the number of variables in @xmath1 $ ] and renders the determination and interpretation of the functional dependence far simpler . in sec . [ sec : wex ] we present and discuss exact results for the interaction energy @xmath1 $ ] of the dimerized hubbard model on representative finite and infinite chains . these are obtained by solving eq . ( [ eq : evgen ] ) using accurate numerical methods . the dependence of the interaction energy on the alternating nn density - matrix elements @xmath86 and @xmath87 is analyzed . scaling properties are identified within the domain of representability of @xmath0 . on the basis of these results we propose , in sec . [ sec : wdim ] , a simple general approximation to @xmath2 which is useful for practical applications . a first test on the accuracy of this approximation is also provided by comparison with available exact solutions . in the following we consider the dimerized 1d hubbard model which in the usual notation is given by@xcite @xmath88 the nn hopping integrals @xmath9 take two alternating values : @xmath89 for @xmath4 odd and @xmath90 for @xmath4 even . the corresponding interaction - energy functional reads @xmath91 = min \left [ u \sum_l \langle \psi [ \gamma_{ij } ] \|\ ; \hat n_{l\uparrow } \hat n_{l\downarrow } \;\|\psi [ \gamma_{ij } ] \rangle \right ] \ ; , \ ] ] where the minimization is performed with respect to all @xmath12-particle states @xmath92\rangle$ ] satisfying @xmath93 \| \sum_\sigma \hat c_{i \sigma } ^{\dagger}\hat c_{j \sigma } \| \psi [ \gamma_{ij } ] \rangle = \gamma_{ij}$ ] for nn @xmath85 . for repulsive interactions @xmath1 $ ] represents the minimum average number of double occupations corresponding to a given degree of electron delocalization , i.e. , to a given @xmath0 . ( [ eq : evgen ] ) then reduces to @xmath94 this eigenvalue problem can be solved numerically for finite systems with various boundary conditions . to this aim we expand @xmath43 \rangle$ ] in a complete set of basis states @xmath95 which have definite occupation numbers @xmath96 at all orbitals @xmath97 : @xmath98 with @xmath99 or @xmath100 . the values of @xmath96 satisfy the usual conservation of the number of electrons @xmath101 and of the @xmath102 component of the total spin @xmath103 , where @xmath104 . for not too large clusters , the state @xmath105 \rangle$ ] corresponding to the minimum in eq . ( [ eq : whub ] ) the ground state of eq . ( [ eq : evhub]) can be determined by sparse - matrix diagonalization procedures such as the lanczos iterative method.@xcite for large chains , the properties of @xmath105 \rangle$ ] can be calculated using the density - matrix renormalization - group ( dmrg ) method@xcite which allows reliable extrapolations to the infinite - length limit . finally , in the absence of dimerization ( @xmath106 ) translational symmetry implies that all nn @xmath0 are the same , and therefore one may set @xmath107 for all nn @xmath85 . the lowest eigenvalue of eq . ( [ eq : evhub ] ) can then be determined from lieb and wu s exact solution of the 1d hubbard model following the work by shiba.@xcite in fig . [ fig : wniv ] the interaction energy @xmath6 of dimerized hubbard chains is shown in the form of constant - energy curves given by @xmath108 , where @xmath3 for @xmath4 odd and @xmath5 for @xmath4 even are the density - matrix elements or bond orders between nn s . @xmath109 stands for the hartree - fock energy , and @xmath110 is a constant ( @xmath111 ) . results are presented for the @xmath112 site ring and for the infinite 1d chain which were obtained from eq . ( [ eq : evhub ] ) by using lanczos - diagonalization and dmrg methods , respectively.@xcite only positive @xmath86 and @xmath87 are considered since this is the relevant domain when all the hopping integrals have the same sign . in bipartite lattices , like open chains or rings with even @xmath51 , the sign of the nn bond orders can be changed without altering @xmath6 by changing the sign of the local orbitals at one of the sublattices . thus , @xmath113 . moreover , @xmath114 as even an odd sites can be interchanged by a simple translation ( @xmath51 even for rings ) . the domain of definition of @xmath6 is restricted by the pure - state representability of @xmath0 . the axes @xmath115 and @xmath116 in fig . [ fig : wniv ] represent a collection of disconnected dimers or fully dimerized states , while @xmath117 corresponds to non - dimerized states . in between , the degree of dimerization can be measured by the angle @xmath118 . the degree of electron delocalization for each @xmath119 is characterized by @xmath120 , which is bounded by @xmath121 in order that @xmath0 remains pure - state representable . the density - matrix elements along the curve @xmath122 are the largest bond orders that can be achieved on a given lattice and for given @xmath51 and @xmath50 [ @xmath123 . they represent the maximum electron delocalization for each @xmath119 and yield the extremes of the kinetic energy @xmath124 , with different @xmath119 corresponding to different @xmath125 . thus , for @xmath122 the density matrix can be represented by the ground state of the uncorrelated hubbard model for some @xmath125 ( @xmath126 ) . in the absence of degeneracy the underlying electronic state @xmath127 is a single slater determinant and @xmath128 . consequently , the upper bound for @xmath129 coincides with the @xmath130 curve in fig . [ fig : wniv ] . the correlation energy @xmath131 vanishes as expected in the fully delocalized limit . for @xmath126 the minimization of the energy @xmath132 as a function of @xmath0 can be stated in terms of the representability of @xmath0 alone . in this case the equilibrium condition yielding @xmath35 is achieved at the borders of the domain of representability , more precisely , when the normal to the curve @xmath122 is parallel to @xmath133 . concerning the lower bound @xmath134 , one should first note that as @xmath129 decreases , @xmath135 , it is possible to construct correlated states @xmath136 \rangle$ ] having increasingly localized electrons . charge fluctuations can then be reduced more efficiently for smaller @xmath129 , and therefore the coulomb interaction energy decreases with decreasing @xmath129 [ see eq . ( [ eq : whub ] ) and fig . [ fig : wniv ] ] . @xmath6 reaches its minimum value @xmath137 in the strongly correlated limit where @xmath138 . for half - band filling this corresponds to a fully localized state having @xmath139 and @xmath140 . however , note that for @xmath141 , @xmath6 reaches @xmath142 already for @xmath143 since partially delocalized states can be found having minimal coulomb repulsion . this is the case for example in a fully - polarized ferromagnetic state . [ fig : wniv ] also provides a qualitative picture of the functional dependence of @xmath6 for dimerized chains . on the one side , for strongly dimerized states ( @xmath144 or @xmath145 ) the constant-@xmath6 curves resemble circumference arcs , the gradient @xmath146 being approximately radial . this type of behavior is most clearly seen for weak or moderate correlations ( @xmath147 ) , while in the localized regime ( @xmath148 ) it holds only for a very limited range of @xmath119 around @xmath149 or @xmath150 . on the other side , for weakly to moderately dimerized states ( @xmath151@xmath152 ) the level curves can be regarded in first approximation as straight lines parallel to @xmath153 . the very weak dependence of @xmath146 on @xmath119 implies that for @xmath154 the ground - state values of @xmath155 and @xmath156 , which result from the minimization of @xmath157 , are very sensitive to the hopping alternation @xmath8 . in fact , significant variations of @xmath119 are necessary until @xmath158 even for @xmath159 . this is particularly notable for weak correlations since the @xmath160 curve is strictly linear for @xmath161 . therefore , a discontinuous change from @xmath162 to @xmath163 is found at @xmath126 and arbitrary small @xmath8 . for @xmath164 , @xmath155 and @xmath156 are continuous functions of @xmath8 , although the dependence on @xmath8 remains very strong for small @xmath8 , as can be inferred from the level curves in the figure . comparing subfigures ( a ) and ( b ) one observes that the results for @xmath112 and @xmath165 are quite similar . the rather rapid convergence with chain length suggests that @xmath166 is not very sensitive to the details of the considered system , even if the minimization constraints in eq . ( [ eq : whub ] ) apply only to nn bond orders . this is of interest for practical applications , as it will be discussed below . in fig . [ fig : wgamma ] the interaction energy @xmath6 is shown as a function of @xmath129 for representative values of @xmath119 , including in particular the fully - dimerized ( @xmath167 ) and non - dimerized ( @xmath168 ) cases . despite the quantitative differences among the various @xmath119 , several qualitative properties are shared by all the curves : ( i ) as already discussed , the domain of representability of @xmath129 is bound for each @xmath119 by the bond order @xmath169 in the uncorrelated limit . @xmath170 decreases monotonously with increasing @xmath119 for @xmath171 showing that a compromise between @xmath86 and @xmath87 is made when the two bonds are active . this is an important contribution to the @xmath119-dependence of @xmath6 . ( ii ) in the delocalized limit , @xmath172 for all @xmath119 , since the electronic state yielding the largest @xmath129 is a single slater determinant . moreover , one observes that @xmath173 diverges at @xmath174 for all @xmath119 . this is a necessary condition in order that the ground - state density matrix satisfies @xmath175 for arbitrary small @xmath164 , as expected from perturbation theory . ( iii ) starting from @xmath174 , @xmath6 decreases with decreasing @xmath129 reaching its lowest possible value @xmath176 for @xmath177 ( @xmath178 ) . the decrease of @xmath6 with decreasing @xmath129 means that the reduction of the coulomb energy due to correlations is done at the expense of kinetic energy or electron delocalization . ( iv ) in the limit of small @xmath129 one observes that @xmath179 . therefore , for @xmath180 , @xmath181 and @xmath182 , a well known result in the heisenberg limit of the half - filled hubbard model.@xcite in order to compare the @xmath129-dependence of @xmath6 for different @xmath119 and to analyze its scaling behavior it is useful to bring the domains of representability for different @xmath119 to a common range by considering @xmath183 as a function of @xmath184 , as displayed in fig . [ fig : wgren ] . in this form the results for different @xmath119 appear as remarkably similar , showing that the largest part of the dependence of @xmath6 on the ratio @xmath185 comes from the domain of representability of @xmath0 given by its upper bound @xmath169 [ @xmath139 for half - band filling ] . an analogous scaling behavior has been found in previous numerical studies of @xmath20 of non - dimerized hubbard models , where @xmath86 refers to the nn density - matrix element.@xcite in this case one observes that @xmath20 depends weakly on system size @xmath51 , band - filling @xmath186 , and lattice structure , if @xmath6 is measured in units of the hartree - fock energy @xmath187 and if @xmath86 is scaled within the relevant domain of representability @xmath188 $ ] . in the present context , [ fig : wgren ] implies that the change in @xmath6 associated to a given change in the degree of delocalization @xmath184 can be regarded as nearly independent of @xmath119 and system size . a good general approximation to @xmath189 can then be obtained by applying such a scaling to the functional dependence extracted from a simple reference system . an appropriate choice is provided by the fully - dimerized chain corresponding to @xmath167 , which can be worked out analytically . in this case the system consists of a collection of dimers and the exact interaction energy reads @xmath190 scaling the functional dependence of the dimer interaction energy to the @xmath119-dependent domain of representability one obtains @xmath191 ^ 2 } \right ) \ ; , \ ] ] which we propose as approximation to @xmath6 for dimerized systems . notice that @xmath192 preserves the previous general properties ( i)(iv ) and that it is of course exact for @xmath193 [ @xmath194 . in practice , the system specific function @xmath169 can be easily obtained by integration of the single - particle spectrum . it is important to remark that the density matrices @xmath0 involved in the approximate functional @xmath195 are pure - state @xmath12-representable . ( [ eq : wdim ] ) applies to the @xmath0 obtained by scaling the off - diagonal elements of the density matrices @xmath196 that derive from uncorrelated states @xmath197 having @xmath50 electrons on @xmath51 sites , and a uniform density distribution @xmath198 . in other terms , @xmath0 has the form @xmath199 with @xmath111 for all @xmath200 , and @xmath201 for all @xmath4 . in order to show the pure - state representability of @xmath0 , we consider two normalized @xmath12-particles states @xmath202 and @xmath203 satisfying @xmath204 and @xmath205 having different defined total spins @xmath206 or @xmath207 , or by superpositions of pure-@xmath206 or pure-@xmath207 states sharing no common eigenvalues . the density - matrix represented by @xmath208 with @xmath209 is then given by @xmath210 , where @xmath211 and @xmath212 are the density matrices corresponding to @xmath213 and @xmath214 . therefore , all the density matrices in the segment defined by @xmath211 and @xmath212 are pure - state @xmath12-representable . the representability of a scaled uncorrelated @xmath196 at half - band filling follows from the previous lemma by taking @xmath215 , which has @xmath216 or @xmath217 , and @xmath214 equal to the fully localized state with one electron per site and maximal @xmath218 , for which @xmath219 for all @xmath220 , and @xmath221 for all @xmath4 . consequently , the @xmath0 in the domain of definition of @xmath195 and the ground - state density matrices @xmath35 derived from it are all pure - state @xmath12-representable . [ fig : wdif ] compares eq . ( [ eq : wdim ] ) with the exact @xmath183 for a @xmath222-site hubbard ring and for the infinite chain . one observes that the proposed approximation follows rather closely the exact results for all @xmath129 and @xmath119 . the largest discrepancies are found for vanishing or moderate dimerization ( e.g. , @xmath223 or @xmath224 ) and relatively large @xmath129 ( @xmath225 ) . in all cases the quantitative differences remain small ( @xmath226 for @xmath224 and @xmath227 for @xmath228 ) which is quite remarkable taking into account the simplicity of the approximation . in the following , eq . ( [ eq : wdim ] ) is applied in the framework of ldft to determine several properties of the dimerized 1d hubbard model . comparison is made with exact results whenever possible in order to assess the performance of the method . in figs . [ fig : e12 ] and [ fig : einf ] the ground - state energy @xmath34 , kinetic energy @xmath229 , and coulomb energy @xmath230 of the 1d hubbard model are given as a function of the coulomb repulsion strength @xmath7 for different hopping alternations @xmath8 . accurate numerical results are also shown which were obtained using the lanczos diagonalization method@xcite for @xmath112 or the dmrg method@xcite for the infinite chain . in the case of the non - dimerized infinite chain , lieb and wu s exact solution@xcite is taken as reference . the results for @xmath112 and @xmath231 are qualitative very similar . @xmath34 increases monotonically with @xmath7 since @xmath232 , vanishing in the limit of @xmath233 . for @xmath234 this is essentially a consequence of the increase of @xmath235 , as @xmath41 and @xmath0 remain very much like in the uncorrelated @xmath126 state . in contrast , for @xmath236 the electrons become increasingly localized , and the increase of @xmath34 results form the increase of @xmath41 which approaches zero as @xmath237 decreases . at the same time @xmath238 tends to zero as charge fluctuations are suppressed ( see figs . [ fig : e12 ] and [ fig : einf ] ) . comparison between ldft and the exact results shows a very good agreement . this concerns not only @xmath34 but also the separate kinetic and coulomb contributions indicating that electron localization and intra - atomic correlations are correctly described for all @xmath7 . moreover , this also shows that the results obtained for the ground - state energy are not the consequence of strong compensations of errors . concerning the accuracy of @xmath41 and @xmath238 one generally observes that a somewhat higher precision is achieved for @xmath41 , which functional dependence is known exactly , as compared to @xmath238 , which derives from an approximation to @xmath6 [ eq . ( [ eq : wdim ] ) ] . for @xmath239 the ldft calculations are nearly indistinguishable from the exact ones ( e.g. , @xmath240 for @xmath241 ) . even the largest quantitative discrepancies , found for the non - dimerized chain at intermediate @xmath7 , are pretty small . for instance , for @xmath106 and @xmath242 we obtain @xmath243 for the 12-site ring and @xmath244 for the infinite chain . comparing figs . [ fig : e12 ] and [ fig : einf ] one observes that the performance of the method is sometimes higher for the 12-site ring than for the infinite chain . for example , fig . [ fig : einf ] shows that @xmath41 ( @xmath238 ) is slightly overestimated ( underestimated ) for @xmath106 and @xmath245@xmath246 , whereas for @xmath112 a much better agreement with the exact result is found ( see fig . [ fig : e12 ] ) . in any case it is important to recall that no artificial symmetry breaking is required to describe correlation - induced localization correctly , as it often occurs in other approaches ( e.g. , mean - field spin - density - wave state ) . moreover , the present calculations remain simple and numerically not demanding , since the minimization of @xmath37 $ ] is performed using analytical expressions for @xmath41 and @xmath6 [ see eqs . ( [ eq : ek ] ) and ( [ eq : wdim ] ) ] . one concludes that ldft , combined with eq . ( [ eq : wdim ] ) as approximation to the interaction energy functional , provides an efficient and correct description of the ground - state properties of the 1d hubbard model in the complete range of interaction strength and dimerization . the charge - excitation or band gap @xmath247 is a property of considerable interest in strongly correlated systems which can can be related to the discontinuity in the derivative of the kinetic and correlation energies per site with respect to the electron density @xmath248 . the determination of @xmath249 constitutes a much more serious challenge than the calculation of ground - state properties like @xmath34 , @xmath41 , and @xmath238 particularly in the framework of a density - functional formalism . results for @xmath249 of the 1d hubbard model are given in figs . [ fig : hubgap12 ] and [ fig : hubgapinf ] as a function of the coulomb repulsion strength @xmath7 for different values of the hopping alternation @xmath8 ( @xmath250 ) . @xmath249 vanishes for @xmath106 and @xmath251 , and increases with increasing @xmath7 or @xmath8 . comparison between ldft and lanczos exact diagonalizations ( @xmath112 ) or the bethe - ansatz solution@xcite ( @xmath165 and @xmath106 ) shows fairly small quantitative discrepancies . in the most difficult non - dimerized case we find @xmath252 for @xmath112 , and @xmath253 for @xmath165 . for small @xmath7 and @xmath106 , @xmath249 is somewhat underestimated for @xmath112 and overestimated for @xmath165 . the latter is mainly due to the fact that eq . ( [ eq : wdim ] ) fails to reproduce the exponential decrease of @xmath249 for @xmath254 ( @xmath165 and @xmath106).@xcite as in previous properties the accuracy improves with increasing @xmath8 . [ fig : hubgap12 ] shows that the ldft results for non - vanishing dimerization and @xmath112 are very close to the exact ones ( @xmath255 already for @xmath256 ) . therefore , one expects that the predictions for @xmath165 and @xmath257 should be reliable . finally , one may note that in the limit of large @xmath7 , the hopping alternation @xmath8 has little effect on the charge gap . as the electrons tend to localize for @xmath258 , @xmath259 where @xmath260 is the energy of the bottom of the single - particle band . the present lattice - density - functional scheme describes correctly the crossover from a band insulator to a mott insulator which occurs in dimerized chains as @xmath7 is varied from the weak - interaction to the strong - interaction regime . a novel density - functional approach to lattice - fermion models has been applied to the dimerized 1d hubbard hamiltonian . in this framework the basic variable is the single - particle density matrix @xmath0 and the key unknown is the interaction - energy functional @xmath1 $ ] . in the present paper we have first investigated the functional dependence of @xmath6 on the density - matrix elements @xmath86 and @xmath87 between nearest neighbors in dimerized chains ( @xmath261 for @xmath4 odd and @xmath262 for @xmath4 even ) . rigorous results for @xmath166 were derived from finite - ring lanczos diagonalizations and from dmrg calculations for the infinite chain . an analysis of these exact results shows that @xmath6 can be appropriately scaled as a function @xmath184 , where @xmath263 and @xmath169 is the largest representable @xmath129 for a given @xmath264 . a simple general approximation to @xmath6 was then proposed which takes advantage of this scaling behavior and which provides with a unified description of correlations from weak to strong coupling regimes . finally , using this approximation , several ground - state properties and the charge - excitation gap of dimerized chains have been determined successfully as a function of coulomb repulsion strength @xmath7 and hopping alternation @xmath8 . the accuracy of the results encourages more or less straightforward applications of the present approach to related problems such as multi - leg ladders or the two - dimensional square lattice with first and second nn hoppings ( @xmath22-@xmath265 hubbard model ) . moreover , the possibility of generalizing the present scaling approximation to an arbitrary number of independent variables @xmath0 deserves to be investigated in detail , since it would open the way to applications in very low symmetry situations including metal clusters and disordered systems . helpful discussions with dr . maurel are gratefully acknowledged . this work has been supported by conacyt ( mexico ) through the project w-8001 ( millennium initiative ) . computer resources were provided by idris ( cnrs , france ) . langreth and m.j . mehl , phys . b * 28 * , 1809 ( 1983 ) ; a.d . becke , phys . a * 38 * , 3098 ( 1988 ) ; j.p . perdew , k. burke , and m. ernzerhof , phys . lett . * 77 * , 3865 ( 1996 ) , and references therein . o. gunnarsson and k. schnhammer , phys . lett . * 56 * , 1968 ( 1986 ) ; a. svane and o. gunnarsson , phys . rev . b * 37 * , 9919 ( 1988 ) ; k. schnhammer , o. gunnarsson , and r.m . noack , _ ibid . _ * 52 * , 2504 ( 1995 ) . see , for instance , h.c . longuet - higgins and l. salem , proc . soc . * a 25 * , 172 ( 1959 ) ; w.p . schrieffer , and a.j . heeger , phys . lett . * 42 * , 1698 ( 1979 ) ; a. karpfen and j. petkov , solid state commun . * 29 * , 251 ( 1979 ) ; c.s . yannoni and t.c . clarke , phys . * 51 * , 1191 ( 1983 ) ; s. suhai , chem . . lett . * 96 * , 619 ( 1983 ) ; m. takahashi and j. paldus , int . j. quantum chem . * 28 * , 459 ( 1985 ) ; h. kahlert , o. leitner , and g. leising , synth . met . * 17 * , 467 ( 1987 ) ; c.m . liegener , j. chem . phys . * 88 * , 6999 ( 1988 ) ; j. ashkenazi , w.e . pickett , h. krakauer , c.s . wang , b.m . klein , and s.r . chubb , phys . * 62 * 2016 ( 1989 ) ; g. knig and g. stollhoff , phys . lett . * 65 * , 1239 ( 1990 ) ; j.q . sun and r. j. bartlett , j. chem . phys . * 104 * 8553 ( 1996 ) ; m.b . lepetit and g.m . pastor , phys . b * 56 * , 4447 ( 1997 ) ; g.l . bendazzoli , s. evangelisti , g. fano , f. ortolani , and l. ziosi , j. chem . phys . * 110 * , 1277 ( 1999 ) . an extension of the definition domain of @xmath37 $ ] to ensemble - representable density matrices @xmath266 is straightforward following the work by valone ( see ref . ensemble density - matrices are written as @xmath267 . in practice , they are much easier to characterize than pure - state density matrices . c. lanczos 45 * , 255 ( 1950 ) ; b.n . parlett , _ the symmetric eigenvalue problem _ , ( prentice - hall , engelwood cliffs , nj , 1980 ) ; j.k . collum and r.a . willoughby , _ lanczos algorithms for large symmetric eigenvalue computations _ , ( birkhauser , boston , 1985 ) , vol . i. constant interaction - energy curves of the one - dimensional ( 1d ) hubbard model as given by @xmath108 , where @xmath268 is the hartree - fock energy and @xmath110 a constant ( @xmath269 ) . the nn density - matrix elements are @xmath270 for @xmath4 odd , and @xmath271 for @xmath4 even . results are given for ( a ) the @xmath112 site ring and ( b ) the infinite chain , both at half - band filling ( @xmath178 ) . @xmath130 corresponds to uncorrelated states and defines the limit of representability of @xmath0 . unless indicated the difference in @xmath110 between contiguous curves is @xmath272 . ] interaction - energy @xmath6 of the 1d hubbard model at half - band filling ( @xmath178 ) as a function of @xmath120 for different values of @xmath273 : ( a ) ring with @xmath112 sites ( b ) infinite chain . the density - matrix elements are @xmath270 for @xmath4 odd , and @xmath271 for @xmath4 even . ] interaction - energy @xmath6 of the 1d hubbard model as a function of @xmath274 for different @xmath273 . @xmath120 and @xmath169 is the largest representable value of @xmath129 for the given @xmath119 , which corresponds to the uncorrelated limit ( @xmath275 , see fig . [ fig : wniv ] ) . results are shown for ( a ) the @xmath112 site ring and ( b ) the infinite chain , both at half - band filling . ] comparison between the exact interaction - energy functional @xmath276 of the hubbard model and the approximation @xmath195 given by eq . ( [ eq : wdim ] ) . results are given for the @xmath112 site ring ( dashed ) and the infinite chain ( solid ) as a function of @xmath120 for different @xmath118 . ] ground - state energy @xmath277 , kinetic energy @xmath229 , and coulomb energy @xmath230 of dimerized hubbard rings with hopping integrals @xmath278 , coulomb interaction @xmath21 , @xmath279 sites , and @xmath178 electrons . the symbols are obtained from exact lanczos diagonalizations@xcite and the solid curves correspond to the present lattice density - functional theory . ] ground - state energy @xmath277 , kinetic energy @xmath229 , and coulomb energy @xmath230 of dimerized infinite hubbard chains with hopping integrals @xmath280 . the symbols are obtained using the density - matrix renormalization group method@xcite and the solid curves correspond to the present lattice density - functional theory ( see fig . [ fig : e12 ] ) . ] charge - excitation gap @xmath249 of dimerized hubbard rings with @xmath112 sites , hopping integrals @xmath280 , and band filling @xmath281 . the symbols refer to exact numerical diagonalizations@xcite and the curves to the present lattice density - functional approach . ] charge - excitation gap @xmath249 of dimerized 1d hubbard chains with hopping integrals @xmath280 and band filling @xmath281 . the crosses refer to exact bethe - anstaz results for @xmath106 ( see also fig . [ fig : hubgap12 ] ) . ]	the dimerized one - dimensional hubbard model is studied in the framework of lattice density - functional theory ( ldft ) . the single - particle density matrix @xmath0 with respect to the lattice sites is considered as basic variable . the corresponding interaction - energy functional @xmath1 $ ] is defined by levy s constrained search . exact numerical results are obtained for @xmath2 where @xmath3 for @xmath4 odd and @xmath5 for @xmath4 even are the nearest - neighbor density - matrix elements along the chain . the domain of representability of @xmath0 and the functional dependence of @xmath2 are analyzed . a simple , explicit approximation to @xmath2 is proposed , which is derived from scaling properties of @xmath6 , exact dimer results , and known limits . using this approximation , ldft is applied to determine ground - state properties and charge - excitation gaps of finite and infinite dimerized chains as a function of the coulomb - repulsion strength @xmath7 and of the alternation @xmath8 of the hopping integrals @xmath9 ( @xmath10 ) . the accuracy of the method is demonstrated by comparison with available exact solutions and accurate numerical calculations . goals and limitations of the present approach are discussed particularly concerning its ability to describe the crossover from weak to strong electron correlations .
on the level of classical field equations , the three dimensional georgi glashow model has two phases : when the mass parameter of the higgs field is negative , the su(2 ) gauge symmetry is broken into u(1 ) , and when it is positive the symmetry is unbroken . the phase of the system can be determined by a local measurement of , say , the scalar field @xmath0 , which vanishes in the symmetric phase but is non - zero in the broken phase . in the broken phase , the field equations have a topologically non - trivial solution , the t hooft polyakov monopole @xcite , whose energy is concentrated around a point - like core . the mass , , the total energy carried by a monopole , decreases when the mass parameter approaches zero from below , and vanishes in the symmetric phase , in the sense that the solution is indistinguishable from the trivial vacuum solution . in many cases , however , we are more interested in the behaviour of the model when fluctuations are taken into account . it is immaterial whether the fluctuations are thermal fluctuations in a classical field theory or quantum fluctuations in a wick rotated ( 2 + 1)dimensional quantum field theory . both of these systems are described by the same partition function , and we shall make no distinction between them . nevertheless , we shall call the treatment based on classical field equations `` semiclassical '' even though it is no more accurate in a classical field theory at a non - zero temperature than it is in a quantum field theory . when fluctuations are present , the above simple picture changes completely . in particular , the ` symmetric ' and ` broken ' phases are believed to be analytically connected to each other @xcite . order parameter candidates that are not gauge invariant , such as @xmath1 , vanish in both phases , and positive definite observables , such as @xmath0 mentioned above , are non - zero in both phases . it would seem natural that a quantity like the mass of a t hooft polyakov monopole , however , should be protected against the effects of the fluctuations by its topology , and that it should therefore serve as an order parameter for the phase transition . if this were the case , the phases could not be analytically connected . one example of this is the abelian higgs model , in which the vortex tension indeed acts as an order parameter @xcite . on the other hand , it is not even obvious that the monopole mass can be given a rigorous definition in a fluctuating theory , because , in general , one can not assume that the field configurations that contribute to the partition function are in some sense close to solutions of classical field equations . this problem was solved in ref . @xcite , however , where the monopole mass was defined as the increase of the free energy when the total magnetic charge of the system is increased by one . furthermore , it was shown how this quantity can be measured in monte carlo simulations . thus , we have a well defined observable , the monopole mass , which could naturally be expected to be zero in the symmetric phase and non - zero in the broken phase , and still the phases are believed to be analytically connected . the purpose of this paper is to explain this apparent paradox . first , we present a calculation based on a simple dilute monopole gas approximation , which predicts that although the monopole free energy is indeed non - zero and roughly equal to its classical value in a system of intermediate volume , it decays to zero at exponentially large volumes . therefore , it should actually vanish everywhere in the thermodynamic limit . this calculation is very similar to polyakov s argument @xcite that the photon has an exponentially small mass in the broken phase . second , we measure the monopole free energy directly in a monte carlo simulation on different volumes using the method developed in ref . we find that the monopole free energy has a volume - independent value in a wide range of lattice sizes , which shows that it corresponds to a localised , point - like object . in agreement with the analytical arguments , however , it eventually starts to decrease , when the volume is large enough . the vanishing of the monopole free energy in the infinite volume limit implies that the monopoles condense . this leads to confinement of electric charge according to the dual superconductor picture @xcite , and our results can therefore be considered as a numerical verification of polyakov s semi - classical argument @xcite that the higgs phase is confining . in particular , since the monopole free energy vanishes in both phases in the infinite volume limit , it does not act as an order parameter , and this resolves the apparent paradox between a smooth crossover and the non - analytic behaviour of the monopole mass in the semiclassical approximation . within the framework of high temperature dimensional reduction @xcite , the three dimensional georgi glashow model is an effective theory for the yang mills theory at high temperatures ( see , for example , @xcite and references therein ) . the phase transition of our model , however , is not related to the deconfinement phase transition of the yang mills theory or qcd . on the other hand , our methods can be generalised to four dimensions in a straightforward way , and they may therefore be applicable also to studying abelian monopoles @xcite in the yang mills theory , in particular whether they condense at the transition point as has been suggested as a possible `` mechanism '' for confinement @xcite . monopole free energies in the yang mills theory have been studied before by several groups @xcite using different techniques . in refs . @xcite fixed boundary conditions were used to create a monopole , but this leads to significant boundary effects . in refs . @xcite a monopole creation operator was used , which lets one measure not only the mass but also correlation functions of the monopole field . with periodic boundary conditions , however , the operator creates not only a monopole , but also an antimonopole somewhere in the system in order to satisfy gauss s law . the advantage of our approach is that the system really has a non - zero magnetic charge , and because translation invariance is preserved , no singularities can arise even near the boundaries of the lattice . the structure of the paper is as follows . we start by discussing the three dimensional georgi glashow model and the lattice definition of its magnetic monopoles in section [ sect:1higgs ] . in section [ sec : expectations ] , we use semi - classical results to motivate our numerical results . we present details of the monte carlo simulations carried out in section [ sec : numerics ] , and the results obtained in section [ sec : results ] . finally we discuss our findings in section [ sec : conclusions ] . in the continuum , the three dimensional georgi glashow model is defined by the lagrangian @xmath2 [ d_i,\phi ] + m^2 \tr \phi^2 + \lambda \left ( \tr \phi^2 \right)^2,\ ] ] where @xmath3 is in the adjoint representation of the su(2 ) gauge group , @xmath4 and @xmath5 $ ] . the partition function of the theory is formally defined as the path integral @xmath6 this can be interpreted as a three dimensional euclidean quantum field theory , or as a classical statistical field theory with the hamiltonian @xmath7 . the coupling constant , @xmath8 , has the dimensions of mass , and we can write the parameters of the theory in terms of dimensionless ratios with the coupling constant @xmath9 and @xmath10 the notation here reflects the fact that the theory is super renormalisable ( in three dimensions ) , and thus only the scalar mass needs a renormalisation counterterm . even this is only necessary up to the two loop level , and its value is known both in the @xmath11 scheme @xcite and in lattice regularisation @xcite . in eq . ( [ equ : ydef ] ) , @xmath12 is the @xmath11 renormalised mass with renormalisation scale @xmath13 . to study this model in a fully non perturbative manner , we formulate the theory in a way that allows numerical solution by monte carlo simulation on a cubic , euclidean lattice consisting of @xmath14 sites , labelled by a triplet of integers @xmath15 . the action is given by @xmath16 , with the lagrangian @xmath17\nonumber \\ & & + \sum_{i}\left\{2a \left[\tr\phi^2(\vec{x } ) - \tr \phi(\vec{x } ) u_i(\vec{x } ) \phi(\vec{x}+\hat{\imath } ) u_i^\dagger(\vec{x } ) \right]\right . \nonumber \\ & & \left . + m^2 a^3 \tr \phi^2(\vec{x } ) + a^3 \lambda \left[\tr \phi^2(\vec{x})\right]^2\right\ } , \label{equ : lagr}\end{aligned}\ ] ] where @xmath18 is the bare lattice mass parameter and @xmath19 is the conventional notation for the bare lattice gauge coupling . we shall treat this lattice theory as an approximation to the continuum one , and therefore we parameterise the theory in terms of the renormalised continuum couplings defined in eqs . ( [ equ : xdef ] ) and ( [ equ : ydef ] ) . we are able to do this , because the relationships between the lattice and continuum couplings are known @xcite , but we shall postpone discussion of them until section [ sec : numerics ] . we shall also express all quantities in continuum units . it is very well known that in the continuum , the field equations have topologically non - trivial solutions , t hooft polyakov monopoles @xcite . they can be characterised by a non - zero winding number of the higgs field at the spatial infinity , @xmath20 where @xmath21 . although @xmath22 itself is gauge invariant , the integrand is not , and therefore it does not have a direct physical interpretation . it can be easily seen , however , that @xmath22 actually corresponds to the magnetic charge associated with the residual u(1 ) gauge invariance . to see this , let us define the magnetic field as @xcite @xmath23.\ ] ] this is a gauge invariant quantity , and agrees with @xmath24 in the unitary gauge @xmath25 . therefore it is indeed the magnetic field associated with the residual u(1 ) symmetry . the corresponding magnetic charge density , @xmath26 , has the following properties : first , because @xmath27 is given by a total derivative , the charge inside a given volume can be expressed as a surface integral . therefore any local deformation of the fields inside the volume can not change the charge inside the volume . second , the magnetic charge inside a given volume is , in fact , @xmath28 and is therefore quantised in units of @xmath29 . these two properties imply that the only way the charge inside a volume can be changed is by moving a magnetic monopole in or out of the volume . in other words , the magnetic charges are topologically stable . what is less well known is that these same properties are also true for the lattice theory . we can define the analogue of eq . ( [ equ : u1field ] ) as @xmath30 here @xmath31 is the lattice u(1 ) field strength tensor , @xmath32 and @xmath33 . in the continuum limit , @xmath34 approaches @xmath35 . if we define the magnetic charge inside a lattice cell as @xmath36,\ ] ] it satisfies the same conditions that guarantee in the continuum the topological stability of magnetic monopoles : the charge is quantised and can be written as a surface integral . these are the same properties that ensure the stability of monopoles in the continuum , and thereby magnetic monopoles are well - defined and absolutely stable objects even in a discrete lattice theory , unlike the instantons of the four - dimensional yang - mills theory . because of the quantisation and stability of magnetic charge , it makes sense to consider ` microcanonical ' partition functions @xmath37 which are restricted to configurations with a given magnetic charge @xmath38 . the full , ` canonical ' partition function is then simply @xmath39 we define the free energy of a given topological sector by @xmath40 and the free energy of a monopole as the free energy difference of sectors @xmath41 and @xmath42 , @xmath43 semiclassically , @xmath44 $ ] , where @xmath45 is the action of the monopole solution with charge @xmath38 . @xmath46 can also be interpreted as the mass of a monopole , and with a slight abuse of language we can generalise into the fully non - perturbative case by defining the monopole ` mass ' @xmath47 by @xmath48 the semiclassical picture would predict that monopoles are massive in the broken phase and massless in the symmetric phase . if this were true , the mass would serve as an order parameter for the phase transition . we measure the monopole free energy following the method of ref . @xcite , which for convenience we briefly review in this section . our strategy is to work on a finite sized system , and impose boundary conditions that force the total magnetic charge of the lattice to be either odd or even , whilst preserving the translation invariance of the system . this is important because translation invariance guarantees the absence of boundary effects . gauss s law rules out periodic boundary conditions , as the total charge is constrained to be zero . however , translation invariance is preserved by any boundary conditions that are periodic up to symmetries of the lagrangian , and in general they allow a non - zero magnetic charge . for instance , ` c periodic boundary conditions ' @xcite @xmath49 are such that the net magnetic charge can be non - zero , but it is constrained to be even @xcite . we shall refer to calculations using such boundary conditions with a subscript ` 0 ' . similarly , if the fields are constrained to behave as @xmath50 on moving around the lattice , the net magnetic charge is odd . we term these ` twisted ( c periodic ) boundary conditions ' , and denote results so obtained by a subscript ` 1 ' . it is easy to see that both sets of boundary conditions are symmetries of the lattice lagrangian . by a gauge transformation , the twisted boundary conditions may be rewritten as ( untwisted ) c periodic boundary conditions everywhere save at the edges of the lattice , where @xmath51 by a suitable redefinition of the fields @xmath52 we can express the twisted boundary conditions as a theory with c periodic boundary conditions everywhere , but with an additional term in the action that depends solely on the gauge fields : @xmath53 where the change in the action is @xmath54.\end{aligned}\ ] ] we emphasise that , because eq . ( [ equ : freeen1 ] ) is equivalent to eq . ( [ equ : bc ] ) with the translation invariant boundary conditions , the choice of coordinates @xmath55 does not affect any observable , and in particular , it does not fix the location of the monopole on the lattice . in physical terms , @xmath56 gives a negative gauge coupling to three orthogonal stacks of plaquettes which are pierced by three mutually intersecting lines on the lattice . these lines are known in the literature as t hooft lines @xcite . a single , open t hooft line creates a pair of dirac monopoles , and has been used to measure their interaction potential in refs . it should be noted , however , that dirac monopoles are rather different from t hooft - polyakov monopoles . they have only half the magnetic charge of the latter , and are singular , non - dynamical objects . in our case , the t hooft lines are closed by the boundary conditions , and therefore they do not create any singularities , but a non - singular t hooft polyakov monopole . the free energy @xmath57 ( or the ` mass ' ) of a monopole is defined by analogy with eq . ( [ equ : fediff ] ) as @xmath58 in the main , however , we shall study the derivative of @xmath57 with respect to the scalar mass parameter , @xmath59 , @xmath60 where @xmath61 is the volume of the system . we know that for sufficiently large @xmath59 in the symmetric phase the free energy of the monopole will go to zero ( at least in the large volume limit ) . if we see the derivative becoming zero , the free energy is at most a constant . in section [ sec : numerics ] we also measure the free energy at a point in the symmetric phase and find it to be consistent with zero . if the derivative is zero all over the symmetric phase , it is reasonable to assume , then , that the free energy itself is becoming zero . we now turn our attention to the semi classical predictions for the t hooft polyakov monopole ( see , for instance , @xcite ) , to which we would like to compare our results from the fully quantised theory . in the broken phase of the theory the scalar field gains a vacuum expectation value ( vev ) @xmath62 the semi classical solution of unit winding number is the t hooft polyakov monopole @xcite , associated with an isolated zero of the scalar field . away from this , the scalar field decays towards its vev , with a characteristic length scale @xmath63 the gauge field simultaneously decays from being su(2 ) to being asymptotically u(1 ) with a length scale @xmath64 we thus have a picture where asymptotically the gauge fields are abelian , save within some extended core whose size is defined by the above length scales where the gauge fields ` unwind ' into the full su(2 ) gauge manifold . we shall find that this scenario remains at least qualitatively valid when quantum corrections are introduced . the mass of this object is , semiclassically , @xmath65 where @xmath66 is the t hooft function . to satisfy the bogomolny lower bound on the mass , @xmath67 also , it is known numerically ( see , for example , @xcite ) that for small @xmath68 , @xmath69 . the derivative of this mass , as in eq . ( [ equ : mderiv ] ) , is @xmath70 assuming that the monopoles are point - like and non - interacting , we can roughly estimate their density to be @xcite @xmath71 which is suppressed by the exponential of the mass . we may also define a mean separation @xmath72 of the monopoles as @xmath73 when @xmath74 , there is a hierarchy between @xmath72 and the fundamental length scales @xmath75 and @xmath76 , and therefore the above assumption of point - like monopoles is valid . this is the semiclassical picture for infinite volume . what we are interested in is what happens in the quantised theory of finite volume , and the interplay of the system size , @xmath77 with the scales above particularly , we wish to know the fate of the monopole mass on large length scales . let us first briefly discuss what happens when the volume of the system is comparable to , or smaller than , the length scales discussed before . the core size of a monopole is given by the correlation length @xmath78 ( we assume for sake of argument that @xmath76 and @xmath75 are comparable ) , and therefore if @xmath79 , there is no room for a monopole in the system . if the system is forced by twisted boundary conditions to contain one monopole , its core will fill the lattice and the whole system will be in the confining phase . on the other hand , the untwisted system is in the higgs phase . the free energy densities of these two phases differ by a certain non - zero amount @xmath80 , which is essentially the latent heat , and as this is the case in the whole volume , we have @xmath81 thus , we can conclude that when @xmath82 , @xmath57 should scale as the volume of the system . when the volume is increased , @xmath83 , the fields start to approach the u(1 ) of the higgs vacuum far from the monopole core of the twisted system . nonetheless , the core will be affected by the boundary conditions , and in general a restriction in the core size by the boundary will lead to an increase in the ( absolute value of the ) free energy and its derivatives . as a rough estimate , if the total non abelian flux inside the monopole core is roughly constant , then the flux density will vary as the inverse of the volume . the total energy of the system would then vary as @xmath84 . ( the figure @xmath85 is a rough limit derived from our results . ) let us then consider a system that is large enough to comfortably accommodate one monopole , but is so small that the fluctuations are not likely to create isolated monopoles ( or , more accurately , well separated monopole antimonopole pairs ) . this is the case when @xmath86 . that is , the entropy action balance is dominated by the action cost , which limits us to the minimum number of monopoles ( and antimonopoles ) required to satisfy the boundary conditions . we expect the free energy difference , eq . ( [ equ : unprog ] ) , to be that between a system of one monopole and an uncharged box . because a monopole is a localised object , the regions far from the monopole core are unaware of the twist in the boundary conditions . @xmath57 only gets a contribution from the monopole core and is therefore independent of the volume . in this case , the identification of @xmath57 with the monopole mass makes sense , and a comparison between the measured values and the semiclassical formul above yields information on the radiative corrections to the semiclassical monopole . as the volume is increased such that @xmath87 , the entropy gain in introducing well separated monopole antimonopole pairs into the vacuum outweighs the action cost and the mean density of topological objects is no longer expected to be the minimum commensurate with the boundary conditions . the free energy required to introduce an extra monopole into the system is now less than the mass of the single monopole , as we demonstrate with a simple model . following ref . @xcite , where a similar effect was discussed in the case of vortices in ( 2 + 1)dimensions , we assume that the density of monopoles is low enough , so that the probability of finding one in any sub - volume of space is independent of whether there are monopoles present elsewhere in the system . in other words , the monopoles are assumed to be point - like or that overlap of the cores is of vanishing measure . as discussed above , this dilute monopole gas approximation is believed to be valid deep in the broken phase . the probability of finding @xmath88 monopoles or antimonopoles ( we do not distinguish ) in a volume , @xmath61 , follows poissonian statistics @xmath89 we apply this to the volume of the whole lattice @xmath90 . we find different normalisation factors for twisted ( @xmath91 , and @xmath92 ) and untwisted ( @xmath93 , and @xmath94 ) boundary conditions . the free energy of the system ( or its derivative ) is extensive and the sum of the free energy of the components for a dilute gas , and considering the entire system we obtain : @xmath95 where @xmath96 was defined in eq . ( [ equ : sgl - deriv ] ) . eq . ( [ equ : multi - deriv ] ) gives the desired plateau for intermediate @xmath61 , but then decays to zero as @xmath97 , beginning once @xmath98 , such that @xmath99 . note that since @xmath100 is simply the typical number of monopoles and antimonopoles created by fluctuations , this result shows that the monopole free energy decays as soon as the fluctuations can create isolated monopoles . crucial in the above calculation is the assumption that the monopoles are non - interacting . although the monopoles at least semiclassically have a long range coulomb interaction , we believe this approximation is justified , because the interaction is non - confining . nevertheless , it is only an approximation , and therefore it must be tested in numerical simulations , as we do in sect . [ sec : results ] . we can also see that the above argument would break down if we tried to apply it to the four dimensional case , where the monopoles are world lines rather than point - like objects . in a euclidean theory , the action of the monopole world line would be proportional to its length , and therefore @xmath101 would vanish exponentially when the limit of infinite time dimension is taken . the same happens for vortices in the three dimensional u(1 ) theory @xcite . in future work we aim to verify that this is also true for vortices in a non abelian theory . on the other hand , if one of the three dimensions is compact as in the ( 2 + 1)dimensional case at a non - zero temperature , @xmath101 is finite , and again the vortex free energy vanishes in the infinite volume limit @xcite . the prediction of the dilute monopole gas approximation that the monopole free energy vanishes in both phases in the infinite volume limit is compatible with the properties the phase diagram of the theory is believed to have . vanishing free energy means that the monopoles condense , and according to the dual superconductor picture @xcite , this gives rise to confinement . indeed , it is known semiclassically that the non - zero monopole density gives the photon a non - zero mass even in the higgs phase @xcite , and this leads to confinement . thus it is natural to assume that the higgs phase is analytically connected to the confining phase @xcite . again , this can only be true if the monopole free energy vanishes in the higgs phase , because otherwise it would act as an order parameter signalling a transition from the higgs to the confining phase . previous studies @xcite have supported the idea of a smooth crossover between the phases , but as they only concentrated on local quantities , they can not be regarded as proofs . for instance , in the three dimensional abelian higgs model , the phase transition can only be seen in practice by measuring non - local observables such as the vortex tension or the photon mass @xcite . in the present case , the predicted non - zero photon mass has not been observed in simulations @xcite . it is clear from the results presented here that the reason for this lies in the very large volumes required . we simulate the georgi glashow model on the lattice via monte carlo importance sampling of the partition functions for both the c periodic and twisted c periodic boundary conditions . updates to the lattice were performed as compound sweeps consisting of one heatbath update to the gauge and scalar fields , followed by two over - relaxation steps to each . measurements were made once per compound sweep . statistical errors were estimated by jack - knife analysis , dividing the data sets into ten bins . for most lattices , the bin size was much longer than the autocorrelation time of the observables , making them independent . this could be seen in an approximate decrease in the statistical errors as @xmath102 as the number of measurements , @xmath103 , was increased . the only lattice on which this was not readily apparent was the @xmath104 , @xmath105 , where the errors did not show such a reduction . this may indicate that , despite considerable computational effort , the ensemble size is still such that the autocorrelation time was comparable to the bin size . error estimates for this ensemble should thus be treated as lower bounds . to illustrate that the twisted boundary conditions ( [ equ : bc ] ) indeed generate a monopole , we show in fig . [ fig : monopole ] the isosurfaces of @xmath106 and the gauge action density @xmath107 in a typical field configuration at @xmath108 , @xmath109 and @xmath110 . the gauge action peaks and @xmath106 dips around the same point , exactly as is expected to happen near the monopole core . because of thermal fluctuations , the isosurfaces are not spherical . we measure the free energy and its derivative with respect to the scalar mass . the former is done via eq . ( [ equ : unprog ] ) . in practice this does not work ; the importance sampling of the theory with untwisted boundary conditions has very small overlap with that of the twisted partition function . this leads to strong sign fluctuations in @xmath56 which leads to a poor convergence of its average through monte carlo simulation . instead , as in refs . @xcite , we can introduce a set of ensembles defined by a real parameter , @xmath111 $ ] : @xmath112 where @xmath113 is the untwisted case , and @xmath114 represents twisted boundary conditions . we then write @xmath115 where the subscript @xmath116 indicates that the expectation value must be measured using eq . ( [ equ : freeeneps ] ) . this gives us the absolute value of @xmath57 , but with the cost that we have to measure expectation values at non - physical values of @xmath116 . we call this the ` method of progressive twisting ' . calculations of the free energy by progressive twisting typically used 10,000 to 20,000 measurements for each of 37 values of the twisting parameter , @xmath116 , which are then numerically integrated . ( for an alternative approach , see ref . @xcite . ) alternatively , the derivative of the free energy , eq . ( [ equ : mderiv ] ) , may be measured directly , which avoids the reweighting problem . we are , however , calculating an intensive quantity as the difference of two approximately extensive numbers . maintaining a constant error on the former demands increasing accuracy in the latter for increasing volume . even allowing for self averaging and the good scaling properties of the simulation algorithm , maintaining comparable precision in the free energy derivative requires cpu time rising as @xmath117 . this limits the results of this study to @xmath118 . calculations of the derivative with respect to the scalar mass used between 200,000 and 500,000 measurements for each of the boundary condition choices . .[tab : prog - params ] lattices used to study the monopole free energy by the method of progressive twisting . [ cols= " < , < , < , < , < " , ] we summarise these estimates of @xmath66 in table [ tab : mass ] , and in fig . [ fig : mass ] where we show a fit to different @xmath59 as per eq . ( [ eqn_sc_mass ] ) . the masses and their derivatives behave much as the semiclassical expectations . similarly the t hooft function , within the limits of our statistical errors , does not appear to differ markedly due to radiative corrections . there is , however , a considerable variation in the data at @xmath119 as we change @xmath120 , and we may worry about systematic effects in our results . the first source of these is discretisation effects . the majority of our estimates are for @xmath121 , and as we have argued above , the residual lattice spacing effects are small here . the variation in @xmath120 in the table is also in part due to a corresponding change in the physical volume of the system , and we may ask whether all our measurements are for ` plateau ' masses uncontaminated by the transient small volume effects . we believe such biases to be small , especially for the @xmath122 . as we vary @xmath59 in fig . [ fig : deriv ] there is a great change in the correlation lengths @xmath123 for fixed volume . that the different _ effective _ volumes considered agree suggests we are indeed seeing the intermediate plateau unaffected by small @xmath124 transients . we are thus confident that our errors on these estimates of @xmath66 are accurate . the joint fit in fig . [ fig : mass ] yields @xmath125 . the large to intermediate system size data for the derivative of the free energy are shown in fig . [ fig : screening ] . for intermediate system size it is clear that the data is well represented by a constant independent of the lattice size , and we use such a hypothesis : @xmath126 where we expect the parameter @xmath127 to be @xmath128 . we show such fits in table [ tab : const - fit ] . our method is to begin with a low upper limit for the fitting range , and to then increase this , including progressively more data in the fit . the @xmath129 per degree of freedom and @xmath130 ( if our fitted form is the correct model , the probability that our data could have arisen as random fluctuations around that model ) remain ( very ) acceptable up to @xmath131 . it is clear that beyond this the fits become unacceptable : the behaviour has changed as a consequence of screening . & & & + 6 & 32 & 10 & 12&431 ( 58 ) & 0&686 & 0&722 + 6 & 36 & 11 & 12&426 ( 58 ) & 1&106 & 0&353 + 6 & 40 & 12 & 12&421 ( 58 ) & 1&704 & 0&066 + 6 & 46 & 13 & 12&419 ( 58 ) & 1&813 & 0&040 + 10 & 32 & 8 & 12&705 ( 82 ) & 0&772 & 0&628 + 10 & 36 & 9 & 12&689 ( 82 ) & 1&229 & 0&272 + 10 & 40 & 10 & 12&677 ( 82 ) & 1&873 & 0&044 + 10 & 46 & 11 & 12&671 ( 82 ) & 1&976 & 0&027 + we can attempt to describe the screening by fitting over a similar range using a fitting ansatz suggested by the dilute gas model : @xmath132 where @xmath127 is as before , and @xmath133 . we show such fits over similar ranges in table [ tab : screen - fit ] . for intermediate @xmath124 the fits are similar to those obtained using just a constant description . as data from larger systems is included , however , we see that the ansatz now remains good . a comparison of the two fits is plotted in fig . [ fig : screening ] . & & & & + 6 & 32 & 9 & 12&499 ( 44 ) & 0&151 ( 93 ) & 0&938 & 0&490 + 6 & 36 & 10 & 12&500 ( 44 ) & 0&187 ( 40 ) & 0&867 & 0&564 + 6 & 40 & 11 & 12&502 ( 44 ) & 0&202 ( 35 ) & 0&826 & 0&614 + 10 & 32 & 7 & 12&445 ( 83 ) & 1&36 ( 106 ) & 0&816 & 0&582 + 10 & 36 & 8 & 12&451 ( 83 ) & 1&82 ( 47 ) & 0&753 & 0&675 + 10 & 40 & 9 & 12&454 ( 83 ) & 1&99 ( 37 ) & 0&722 & 0&689 + 10 & 46 & 10 & 12&449 ( 83 ) & 1&79 ( 35 ) & 0&773 & 0&655 + we may calculate from @xmath134 the mean density of monopoles , @xmath135 , which makes their mean separation @xmath136 or , in lattice units at @xmath137 , @xmath138 . from this it is clear that we have not got the lattice volume necessary to see a complete screening of the free energy at @xmath139 . we can not therefore completely rule out from our data the possibility that @xmath57 remains finite even in the infinite volume limit . as was seen for small system sizes , the plateau values at least are heavily influenced by discretisation effects at @xmath137 . to perform a scaling study of the screening mechanism is beyond our current means . nonetheless , for a demonstration of the mechanism such effects are immaterial and do not affect the qualitative arguments . note also that no attempt has been made to estimate here the systematic errors in the monopole density . to do so would require a comparison of different screening hypotheses and fit functions , something that the data is , unfortunately , not accurate enough to address satisfactorily . in this paper , we have used a fully non - perturbative technique to measure the free energy of a t hooft polyakov monopole in the three dimensional georgi glashow model . this was achieved by simulating systems with two different boundary conditions , both of which are periodic up to symmetries of the lagrangian . this preserves the lattice translation invariance of the system and therefore makes sure there are no boundary effects . we found that in the higgs phase , the free energy reached a constant value at intermediate volumes , which shows that it is associated with a localised object . this is the quantum analogue of the t hooft polyakov monopole . we measure its mass by two different methods , and find it compatible with semiclassical expectations . ` mean field ' application of the classical relations appears successful , and we can make estimates of the quantum corrected t hooft function . our best estimates are @xmath140 and @xmath141 from the derivative of the mass with respect to @xmath59 , and @xmath142 by the method of progressive twisting . these estimates are both self consistent , and in agreement with the classical variation @xmath69 for small @xmath68 @xcite , indicating that radiative corrections are small . when the volume increased above a certain critical value , however , the free energy started to approach zero . this is consistent with an analytical calculation within the dilute monopole gas approximation , which predicts that the free energy vanishes in the infinite volume limit at any values of the couplings . in the dual superconductor picture , the vanishing monopole free energy implies confinement , and therefore our results are numerical evidence for polyakov s prediction that the higgs phase of this theory is confining . furthermore , if the monopole free energy vanishes everywhere , it can not be used as an order parameter , and therefore our results strongly support the conjecture that the confining and higgs phases are analytically connected to one another . neither , of course , can the monopole mass measured from the plateau in the free energy for intermediate system sizes act to distinguish the phases . it is non - zero in the deep higgs phase and zero in the deep symmetric phase . this plateau does not exist , however , everywhere in the phase diagram , notably near the transition line itself . the mean monopole separation there will be comparable to the core size and no plateau would be observed . thus the ` mass ' is ill defined and can not serve as an order parameter . our findings suggest a straightforward generalisation to other cases . in a euclidean formulation in any number of dimensions , any point - like topological defect that has finite action , will always have a non - zero density at any non - zero temperature . this means that these objects always have a zero free energy . an extended topological defect , such as a string or a domain wall , is , however , either a closed loop , surface etc . , in which case it does not contribute to the global properties of the systems , or it has an infinite action . in the latter case , the fluctuations can not generate them , and their free energy remains non - zero even in the infinite volume limit . because the free energy can be used as an order parameter , this suggests that models with extended topological defects always have a true phase transition rather than a smooth crossover . this question will be studied further in a future publication . we would like to thank mikko laine and kari rummukainen for useful discussions . this work was supported by pparc(uk ) and by the esf coslab programme . the computational work was carried out on the u.k . computational cosmology consortium cosmos origin2000 supercomputer . g. t hooft , nucl . b 79 ( 1974 ) 276 . a. m. polyakov , jetp lett . 20 ( 1974 ) 194 . e. fradkin and s. h. shenker , phys . d 19 ( 1979 ) 3682 . s. nadkarni , phys . rev . lett . 60 ( 1988 ) 491 ; nucl . b 334 ( 1990 ) 559 . a. hart , o. philipsen , j. d. stack and m. teper , phys . b 396 ( 1997 ) 217 . k. kajantie , m. laine , k. rummukainen and m. shaposhnikov , nucl . b 503 ( 1997 ) 357 . h. kleinert , lett . nuovo cim . 35 ( 1982 ) 405 . a. kovner , b. rosenstein and d. eliezer , nucl . b 350 ( 1991 ) 325 . k. kajantie , m. laine , t. neuhaus , j. peisa , a. rajantie and k. rummukainen , nucl . b 546 ( 1999 ) 351 . a. c. davis , t. w. b. kibble , a. rajantie and h. shanahan , jhep 0011 ( 2000 ) 010 . a. m. polyakov , nucl . b 120 ( 1977 ) 429 . a. m. polyakov , gauge fields and strings ( harwood , 1987 ) . g. t hooft , in `` high energy physics '' , eps international conference , palermo 1975 , ed . a. zichichi ; s. mandelstam , phys . 23 ( 1976 ) 245 . k. kajantie , m. laine , k. rummukainen and m. shaposhnikov , nucl . b 458 ( 1996 ) 90 . a. hart and o. philipsen , nucl . b 572 ( 2000 ) 243 ; a. hart , m. laine and o. philipsen . b 586 ( 2000 ) 443 ; phys . lett . b 505 ( 2001 ) 141 . g. t hooft , nucl . b 190 ( 1981 ) 455 . j. smit and a. j. van der sijs , nucl . b 422 ( 1994 ) 349 . l. del debbio , a. di giacomo and g. paffuti , phys . b 349 ( 1995 ) 513 ; a. di giacomo , b. lucini , l. montesi and g. paffuti , phys . d 61 ( 2000 ) 034503 . j. frhlich and p. a. marchetti , nucl . b 551 ( 1999 ) 770 . p. cea and l. cosmai , phys . d 62 ( 2000 ) 094510 . k. farakos , k. kajantie , k. rummukainen and m. shaposhnikov , nucl . b 442 ( 1995 ) 317 . m. laine , nucl . b 451 ( 1995 ) 484 . m. laine and a. rajantie , nucl . b 513 ( 1998 ) 471 . a. s. kronfeld and u. j. wiese , nucl . b 357 ( 1991 ) 521 ; nucl . b 401 ( 1993 ) 190 . g. t hooft , nucl . b 138 ( 1978 ) 1 . t. g. kovacs and e. t. tomboulis , phys . 85 ( 2000 ) 704 . c. hoelbling , c. rebbi and v. a. rubakov , phys . d 63 ( 2001 ) 034506 . a. hart , b. lucini , z. schram and m. teper , jhep 11 ( 2000 ) 043 . de forcrand , m. delia and m. pepe , phys . lett . 86 ( 2001 ) 1438 . m. n. chernodub , f. v. gubarev , m. i. polikarpov and v. i. zakharov , phys . b 514 ( 2001 ) 88 . p. goddard and d. i. olive , rept . phys . 41 ( 1978 ) 1357 .	we study the three - dimensional georgi glashow model to demonstrate how magnetic monopoles can be studied fully non - perturbatively in lattice monte carlo simulations , without any assumptions about the smoothness of the field configurations . we examine the apparent contradiction between the conjectured analytic connection of the ` broken ' and ` symmetric ' phases , and the interpretation of the mass ( , the free energy ) of the fully quantised t hooft polyakov monopole as an order parameter to distinguish the phases . we use monte carlo simulations to measure the monopole free energy and its first derivative with respect to the scalar mass . on small volumes we compare this to semi classical predictions for the monopole . on large volumes we show that the free energy is screened to zero , signalling the formation of a confining monopole condensate . this screening does not allow the monopole mass to be interpreted as an order parameter , resolving the paradox .
near a supermassive black hole ( smbh ) tidal disruptions occur during close encounters when a plunging star on an eccentric orbit reaches the critical ( tidal ) radius @xmath0 cm , where @xmath1 and @xmath2 denote the mass and the radius of the satellite star , @xmath3 is the smbh mass @xcite . stars approach the event horizon , and at a certain moment they become disrupted by tidal forces of the smbh , producing a bright flash of intense radiation that illuminates the surrounding interstellar medium and a temporary accretion disc or a ring formed by the debris @xcite . according to the standard picture of galactic nuclei , we can imagine a nuclear cluster around a galactic core as a system consisting of a central smbh , an accretion disc and a dense stellar cluster , possibly of a flattened shape in its inner region @xcite . the fall - back rate of the remnant debris onto smbh is expected to be influenced by relativistic effects @xcite . the authors of ref . @xcite demonstrate that the photoionised debris of a tidally disrupted hb star can account for the emission lines observed in some optical spectra . in their case , the super - eddington phase lasts about one to two hundred years ; reproducing the line ratios requires an intermediate - mass black hole of @xmath4 . various characteristics of tdes depend strongly on the stellar type @xcite , although the gradual decay of the light curve adopts a generic profile that is determined by the viscous processes . the emission of an x - ray irradiated flow , known as the reflection spectrum , can be expected , including a superposition of the continuum emission and spectral lines , including the prominent fluorescent emission lines of iron that have the rest energy around 67 kev @xcite . in another context of x - ray emission lines , a method for o stars to determine the shock - heating rate by instabilities in their radiation - driven winds has been recently developed @xcite . for this paper , we just remind the reader that the relativistically smeared spectral line emission from a black - hole accretion disc in a few kev band may be intrinsically narrow and unresolved in energy ; it is the observed profile that becomes broadened and skewed ( by a combination of doppler and light - bending effects ) when integrated over the azimuthal extent of accretion rings , and generally red - shifted by strong gravity of the central black hole . one of the most constraining x - ray spectral information on a tidal disruption event ( tde ) has resulted from the campaign on a quiescent galaxy sdss1201 + 30 @xcite . this is most likely an object without a prior accretion disc . the x - ray spectra are very soft , and can hardly be explained with standard accretion disk models . the strong variability was seen in the light curve and related to clumpy accretion with a combination of flaring and absorption events . x - ray photons from tde can illuminate , ionise and perturb the gaseous material infalling from the fresh accretion flow . in ref . @xcite we propose that the relativistic iron line from the tde irradiated inflow can provide a unique way to determine parameters of the system , namely , the dimensionless black - hole spin @xmath5 ( @xmath6 ) , the angle of the observer @xmath7 ( inclination , @xmath8 deg ; @xmath9 corresponds to pole - on view along the rotation axis ) , and the expansion velocity @xmath10 of ionisation front , which propagates outwards at velocity close to the speed of light and modulates the ionisation parameter of the medium . hence , it affects also the iron - line emissivity . according to the model , the line emission is triggered and modulated by the same , virtually instant tde event . unlike the case of continuously variable x - ray spectrum of active galactic nuclei @xcite , a quiescent source would be preferred , so that the transient relativistic line can be revealed . in ref . @xcite we modelled the structure of the nuclear star cluster that is expected to arise as a result of the mutual interaction of main components of the galactic nucleus , i.e. , a supermassive black hole , stars , and a gaseous / dusty torus that helps to bring stars towards the tidal disruption . an interesting conclusion from that paper concerns the formation of a flattened ( oblate ) stellar subsystem in the inner region surrounding the central smbh @xcite , which is a reminiscence on much smaller scales of the structure originally reported on the kiloparsec scale of the bulge of the milky way @xcite . in the other words , tde from such a stellar population can be expected to occur predominantly in the equatorial plane of the system . therefore , also the irradiation flash should arise with a higher chance within the plane , i.e. , at large inclination angle with respect to a distant observer . in our scenario for tde , the flattened stellar system is associated with the effect of hydrodynamical dissipation of the stellar orbits by the dusty / gaseous torus . further down , at very small radii near the horizon , also gravitational radiation can play a role . moreover , a subset of stars on eccentric orbits is brought close to the central black hole where they plunge quickly below the tidal radius . studying such a tde would shed light on the accretion physics and the stellar dynamics in galactic nuclei , as well as the role of tidal disruptions for feeding and growth of smbhs . we put forward a possibility that the source of intense energy deposition occurs near the inner edge of the accretion disc , which resembles a radially narrow ring that gradually spreads by viscous forces and eventually disappears once all material is accreted . basic form of the relativistic line from a narrow ring were investigated in ref . following tde , the spectral line is modulated via changing ionisation state of a remnant accretion disc that is created from the debris . this remnant accretion ring can be identified with a structure that has been proposed to arise around the circularisation radius at transient accretion events with low ( sub - keplerian ) angular momentum @xcite . . as the ionisation front expands through the infalling remnant disc , the changing ionisation state is revealed by the corresponding change of the iron line rest energy and intensity , until the emission feature disappears when the complete ionisation is reached.,scaledwidth=60.0% ] a non - negligible radial infall velocity of the accreted material is likely , and so the ionisation front moves radially outward through the material of the remnant disc at @xmath11 that is not exceeding the speed of light ( although it should be very close to it ) . we can set @xmath12 as one of free model parameters . the corresponding mass fall - back rate proceeds as a characteristic power - law profile @xcite , @xmath13 ( @xmath14 ) , which can significantly exceed the eddington accretion rate for a period of weeks to years for the black hole mass @xmath15 @xcite . the hypothesis about an enhanced capture rate and tidal disruptions of stars in the equatorial plane is supported also by simulations of the structure flattening in the inner regions of the nuclear cluster . in ref . @xcite a scenario was discussed based on a combination of simultaneous gravitational and hydrodynamical effects of the gaseous environment on orbiting stars , which are assumed to lead to an oblate , disc - like configuration . while the direct star - disc hydrodynamical interaction causes a continuous dissipation of the stellar orbital energy ( and it is anyway very small in the milky way s central regions ) , gravity of the stellar ring , a self - gravitating accretion torus , or a flattened nuclear star - cluster all can lead to periodical variations of the orbital elements . an example of this evolution is plotted in figure [ fig2 ] , where we show the long - term orbital changes due to both the hydrodynamical and gravitational influence of the disc . as an example , parameters of that simulation were set to be consistent with the values reported for s2 star in the galactic center . alternatively , tidal disruption of stars by supermassive central black holes from dense star clusters has been modelled by high - accuracy direct n - body simulations @xcite . ) .,scaledwidth=99.0% ] we note that in the x - ray and optical / uv bands , almost two dozens candidate tdes have been already reported @xcite . these events are characterized by the thermal emission with temperature of @xmath16@xmath17 k , and the peak bolometric luminosity about @xmath18@xmath19 . for the events with good coverage during the decay , the flux decline was found consistent with a power - law ( index @xmath20 ) , as predicted for canonical tdes . compared with the quiescence state , the flux can increase by a factor of @xmath21 during outbursts . detections of the fluorescent iron line from tde are only tentative so far @xcite , however , finding this spectral signature should provide very valuable information . the expected properties of the reflection line depend sensitively on the ionisation state of the irradiated material @xcite . since in tdes the continuum flux can vary by a large amplitude , we need to take also the variation of the ionisation parameter @xmath22 into consideration ( see figure [ fig1 ] ) . for @xmath23 smbh , the peak mass fall - back rate can reach @xmath24 , corresponding to luminosity of @xmath25 ( assuming the radiation efficiency of @xmath26 ) . therefore , the rise of the illumination is significant by orders of magnitude compared to the quiescent state of a dormant nucleus . gravitational effects act on the spectral features from the remnant tde accretion disc by smearing the spectral features and moving the observed energy centroid across energy bins . in this way gravity exerts the influence on the ultimate form of the spectrum @xcite . the reprocessed radiation reaches the observer from different regions of the system . furthermore , as strong - gravity plays a crucial role , photons may even follow multiple separate paths , joining each other at the observer at different moments . individual rays experience unequal time lags for purely geometrical reasons and for relativistic time - dilation . time delay from a tde flare to the moment of arrival of the observed iron line signal consists of two components : the expansion time @xmath27 , and the time interval from the disc to the observer , @xmath28 . at large radius where the spacetime is flat to a good precision , @xmath29 , where @xmath30 is the azimuthal angle on the disc plane @xcite . as @xmath10 is of order unity ( i.e. , comparable to the speed of light ) , the two quantities become comparable , and the delay interval is dominated by the longer one . at large inclination angles , the photons from the disc located in front of the black hole reach the observer first , then those from disc near the inner edge , and at last those from disc behind black hole . in the cases of high expansion velocity combined with high inclination angle , a `` nose '' occurs ahead of the rings main signal @xcite . the length of the nose increases with the inclination angle and with the expansion velocity . the line emission ceases once the accretion flow becomes highly ionised to a larger distance . on a phenomenological level of our model , the intrinsic emissivity of the remnant accretion ring and the expansion velocity of the ionisation front are two degrees of freedom that allow us to capture the lamp - post scheme as well as the tde - induced illumination in a common scenario . we note that the remnant accretion rings resulting from a tde event are likely lacking axial symmetry . therefore , as a next step to explore more realistic situations , spectral line profiles from elliptically shaped and tilted structures need to be taken into account @xcite . we conclude by stating that the proposed idea poses an observational challenge ( because the iron line flux is expected to be only a weak and variable components of the x - ray signal from tidal - disruption events ) , nevertheless , it suggests a promising opportunity to verify the parameters of central black holes by an independent method . moreover , it offers and interesting complement to the standard scenario for the origin of relativistically broadened spectral lines . in the context of galactic centre , let us remark that the 67 kev emission of iron has been extensively studied @xcite and the light - echo effect reported . however , the detection concerns a wider region ( molecular clouds within the central molecular zone surrounding the smbh ) than the immediate vicinity of the black - hole horizon , which we imagine here . while in the lamp - post model the irradiating source is ( usually ) considered to be ( almost ) axially symmetric and located around the rotation axis , in the present scheme the primary excitation is more likely to occur from the equatorial region near the innermost stable circular orbit , commonly denoted as isco , which is about the minimum radius to which the accretion discs can extend . a convincing case of such a transient relativistic spectral feature from tde is still to be found in x - rays . the authors highly appreciate the hospitality of conference organisers in prague . vk and dk thank the czech ministry of education , youth and sports project lh14049 , titled `` spectral and timing properties of cosmic black holes '' , that has been aimed to support the international collaboration in astrophysics . vk and md acknowledge also the european union 7th framework programme no . 312789 `` stronggravity '' . the research of wy and wz is supported in part by the national natural science foundation of china under grants no . 11073043 and no . 11333005 . 99 bu de - fu , feng yuan 2014 , mnras * 442 * , 917 cannizzo j. k. , lee hyung mok , goodman j. 1990 , apj , * 351 * , 38 chang h .- y . , choi c .- s . 2002 , journal of the korean astronomical society * 35 * , 123 cheng r. m. , bogdanovic t. 2014 , phys . rev . d , in press ( arxiv:1407.3266 ) clausen d. , sigurdsson s. , eracleous m. , irwin j. a. 2012 , mnras * 424 * , 1268 cohen d. h. , li zequn , gayley k. g. , owocki s. p. , sundqvist j. o. , petit v. , leutenegger m. a. 2014 , mnras , in press ( arxiv:1409.0856 ) czerny b. , kunneriath d. , karas , v. , das t. k. 2013 , a&a * 555 * , a97 eracleous m. , livio m. , halpern j. p. , storchi - bergmann t. 1995 , apj * 438 * , 610 esquej p. , saxton r. d. , komossa s. , read a. m. 2012 , in _ tidal disruption events and agn outbursts _ ( madrid , spain ) , edited by r. saxton & s. komossa ( epj web of conferences ) , vol . * 39 * , i d . 02004 evans c. r. , kochanek c. s. 1989 , apjl * 346 * , l13 fragile p. c. , miller w. a. , vandernoot e. 2005 , apj * 635 * , 157 gezari s. , chornock r. , rest a. , huber m. e. , forster k. , berger e. et al . 2012 , nature * 485 * , 217 greiner j. , schwarz r. , zharikov s. , orio m. 2000 , a&a * 362 * , l25 hayasaki k. , stone n. , loeb a. 2013 , mnras * 434 * , 909 just a. , yurin d. , makukov m. , berczik p. , omarov c. , spurzem r. , vilkoviskij e. y. 2012 , apj * 758 * , i d . 51 karas v. 2006 , an * 327 * , 961 karas v. , czerny b. , abrassart a. , abramowicz m. a. 2000 , mnras * 318 * , 547 karas v. , martocchia a. , ubr l. 2001 , pasj * 53 * , 189 komossa s. , greiner j. 1999 , a&a * 349 * , l45 kunneriath d. , schdel r. , stolovy s. , feldmeier a. 2014 , in _ the galactic center : feeding and feedback in a normal galactic nucleus _ ( santa fe , nm ) , proceedings of the international astronomical union symposium 303 , edited by l. sjouwerman , j. ott , & c. lang ( cambridge university press , cambridge ) , p. 228 launhardt r. , zylka r. , mezger p. g. 2002 , a&a * 384 * , 112 luminet j .- , marck j .- a . 1985 , mnras * 212 * , 57 macleod m. , guillochon j. , ramirez - ruiz e. 2012 , apj , * 757 * , i d . 134 ponti g. , terrier r. , goldwurm a. , belanger g. , trap g. 2010 , apj * 714 * , 732 rees m. j. 1988 , nature * 333 * , 523 ross r. r. , fabian a. c. 1993 , mnras * 261 * , 74 saxton r. d. , read a. m. , esquej p. , komossa s. , dougherty s. , rodriguez - pascual p. , barrado d. 2012 , a&a * 541 * , a106 schdel r. , feldmeier a. , kunneriath d. , stolovy s. , neumayer n. , amaro - seoane p. , nishiyama s. 2014 , a&a * 566 * , a47 sochora v. , karas v. , svoboda j. , doviak m. 2011 , mnras , * 418 * , 276 strubbe l. e. , quataert e. 2009 , mnras * 400 * , 2070 ubr l. , karas v. 2005 , a&a * 433 * , 405 ubr l. , karas v. 2005 , in _ black holes and neutron stars _ , proceedings of ragtime 6/7 ( opava , czech republic ) , edited by s. hledk and z. stuchlk ( silesian university in opava ) , p. 281 ubr l. , karas v. , hur j .- 2004 , mnras * 354 * , 1177 vilkoviskij e. y. , czerny b. 2002 , a&a * 387 * , 804 wang q. d. , dong h. , lang c. 2006 , mnras * 371 * , 38 zhang w. et al . 2014 , to be submitted zhong s. , berczik p. , spurzem r. 2014 , apj * 792 * , i d . 137	we discuss a possibility that a tidal disruption event near a dormant supermassive black hole ( smbh ) can give rise to spectral features of iron in 67 kev x - ray signal : a relativistic line profile emerges from debris illuminated and ionised by an intense flash produced from the destroyed star . this could provide a unique way to determine parameters of the system . we consider a model where the nuclear stellar population acquires an oblate shape ( i.e. , a flattened distribution ) in the inner region near a supermassive black hole , and also the primary irradiation flare is expected to occur more likely near the equatorial plane , co - planar with the infalling material . this suggests that the reprocessing of primary x - rays results in a transient profile that should be relevant for tidal - disruption events ( tde ) in otherwise under - luminous ( inactive ) galactic nuclei , i.e. with no prior accretion disc . resonance mechanisms of the stellar motion can increase the orbital eccentricity for some stars in the nuclear cluster and help to bring them close to the tidal radius , where they can give rise to tdes . the proposed scenario appears to be distinguishably different from the standard scheme of lamp - post model for the origin of the relativistic line in active galaxies , where the source is thought to be located predominantly near the symmetry axis . accretion : accretion discs black - hole physics galaxies : nuclei tidal disruption events
high t@xmath4 cuprate superconductors differ from their low t@xmath4 counterparts in their extreme type ii character with a ginzburg - landau parameter @xmath5 corresponding to extremely short coherence length @xmath6 nm and high upper critical fields @xmath7 t. the small abrikosov vortex core size of order @xmath8 results in a series of bound states with energies situated in the gap whose mean energy spacing is of order @xmath9 where @xmath10 is the superconductor gap energy ( modulus of the order parameter ) , @xmath11 the fermi energy and @xmath12 the effective mass . the core structure can affect the mobility for translational motion of the vortices by modifying the spectral properties of the electron momentum transfer . dissipation in type ii superconductors in magnetic field @xmath13 in slowly varying electromagnetic fields is dominated by the dynamics of abrikosov vortices @xcite . material details of the superconductor enter mainly via the friction coefficient that gives the vortex velocity in terms of the driving magnus - lorentz force . the friction is controlled by the spectral density and relaxation properties of the low - lying core states , i.e. , quasiparticle states localized to the vortex core . in dirty superconductors , where the mean free path @xmath14 , the friction coefficient is related to the normal state resistivity by the bardeen - stephen ( bs ) rule @xcite : @xmath15 where @xmath16 is the flux flow resistivity arising from vortex motion in the absence of pinning , @xmath17 is the normal state resistivity , @xmath18 is the magnetic field , @xmath19 the upper critical field and @xmath20 . a more careful examination of the core states allowed the extension of eq . ( [ eq : bs ] ) with @xmath21 to the moderately clean limit @xmath22 as well @xcite . the bs law has been experimentally confirmed for a broad range of conventional ( @xmath23-wave ) superconductors @xcite . in unconventional superconductors with gap nodes , bcs theory suggests a high density of core states at the fermi energy . nevertheless calculation @xcite of the flux - flow resistivity revealed that as long as one stays in the moderately clean limit , eq . ( [ eq : bs ] ) remains valid with a possible enhancement of @xmath24 but still of order @xmath25 . recent measurements @xcite on several anisotropic non - high-@xmath26 superconductors that most likely exhibit gap nodes confirm eq . ( [ eq : bs ] ) with a moderately enhanced @xmath27 the problem of low - temperature vortex dynamics in bscco ( bi@xmath28sr@xmath29cacu@xmath28o@xmath30 is particularly interesting in this respect in the light of the recent results on the structure of the vortices in this material . scanning tunneling microscope ( stm ) spectroscopy @xcite revealed that the zero - energy peak in the density of core states is missing . this , together with the results of inelastic neutron scattering @xcite and nmr @xcite experiments , led to the suggestion that some concurrent , non superconducting , order exists at the vortex cores when the superconducting order parameter is suppressed @xcite . the idea received direct support from stm spectroscopy@xcite where a periodic modulation of the local density of electronic states around the vortex cores was observed . since the decay length of this modulation is much longer than the superconducting coherence length , not only the structure of the core , but also the structure of the flow field for vortex transport should be different from conventional abrikosov vortices with probable consequences for the velocity - force relation . the experimental situation in the high-@xmath26 superconductors , and in particular in bscco , is not clear . low - frequency transport measurements @xcite in the vortex liquid phase close to the critical temperature @xmath26 are in reasonable agreement with the bs law . microwave and millimeter wave impedance measurements @xcite at low temperatures , on the other hand , indicate a large and nearly @xmath18-independent dissipation in a broad field range . to clarify the situation and to see if these very special aspects of the vortex structure influence the dynamics , we have performed a systematic investigation of the flux - flow resistance in bscco single crystals by measuring the @xmath31-@xmath32 plane voltage - current ( @xmath33-@xmath34 ) characteristics up to currents well above the threshold current for dissipation in @xmath3-directed magnetic field . the free flux - flow resistance as measured on an @xmath35 face can be approximately described as @xmath36 for low fields followed by saturation ( becomes field independent ) above about 1 t in the low - temperature vortex solid phase . as the saturated value corresponds quite well with the extrapolation of the normal resistance to low temperatures , it might be naively interpreted as reflecting @xmath37 in eq . ( [ eq : bs ] ) , at least at 1 t , indicating that vortices move at up to two orders of magnitude faster than predicted by bs law in low fields . although it is known that vortex velocities much larger than the bs value may result from a nonlinear instability at high vortex velocities @xcite , not only does its onset in bscco films@xcite occur at current densities about two orders of magnitude higher than investigated in this study , the characteristic signature @xcite of a nonlinear runaway in the @xmath33-@xmath34 curve is absent in our case . measurements were made on nine single crystals from three different batches of bi@xmath28sr@xmath28cacu@xmath28o@xmath38 fabricated by a melt cooling technique @xcite . the typical dimensions of the crystals were @xmath39 mm@xmath40 with the shortest dimension corresponding to the poorly conducting @xmath3 axis . most of the crystals have been close to optimal doping with a resistance - determined critical temperature @xmath41 k and transition width about 2 k in zero field . the diamagnetism in a @xmath42oe field set in in a uniformly progressive way below @xmath26 to near @xmath43 at low temperature . two underdoped crystals exhibited a resistive @xmath44 k and @xmath45 k with width 6 - 8 k and slow initial diamagnetism onset to about @xmath46k before proceeding to about @xmath47 flux exclusion at low temperature . the resistance measurements were made in the usual four point configuuration on an @xmath35 face of the crystals . the contacts to inject and withdraw the current are either on the top @xmath31-@xmath32 face or encompass the ends of the crystal extending to both opposing @xmath31-@xmath32 faces . the features reported here are common to both geometries . four more contacts on the same @xmath31-@xmath32 face as the current injection serve to measure the voltage parallel and transverse to the current flow . the contacts are made by bonding 25 @xmath48 gold wires with silver epoxy fired at @xmath49 k in an oxygen atmosphere resulting in contact resistances of less than 3 @xmath50 for the current contacts . the sample is inserted in the bore of a superconducting magnet with the field along the @xmath3 direction . to measure at the high currents required for the free flux flow regime without significant joule heating , we apply short ( typically 50 @xmath51s or less ) current pulses of isosceles triangular shape with a repetition period of 0.2 to 1 s. further technical details are given and the issue of joule heating for the such experimental conditions is analyzed in ref . @xcite with the conclusion that the temperature change in the area between the voltage contacts is negligible for the duration of the pulse . typical @xmath33-@xmath34 characteristics measured at different temperatures and in a field of 3 t are shown in fig . [ fig : iv ] . at temperatures below the vanishing of zero current resistance , which we interpret as the freezing of the vortex system ( in @xmath52 t this happens at @xmath53 k ) , dissipation sets in abruptly beyond a sharply defined threshold current @xmath54 . usually , but not always @xcite , a linear segment appears above the threshold current ( see insert ) the slope of which we denote by @xmath55 . at higher currents , the differential resistance increases further and saturates at a value @xmath56 in the high - current limit . the high - current linear segment of the @xmath33-@xmath34 characteristic extrapolates to finite current @xmath57 at zero voltage . above the melting temperature , a finite differential resistance @xmath58 is observed in the @xmath59 limit . the @xmath33-@xmath34 curve , however , is still nonlinear and @xmath60 is finite up to a temperature @xmath61 situated between the melting temperature and the critical temperature @xmath26 . t at temperatures ( from left to right ) 60 , 45 , 38 , 19 , 11 , 7 , and 5 k. inset : magnified view of the low - current region . the dashed lines are linear fits . ( see text for the definition of @xmath60 , @xmath54 , @xmath62 , and @xmath55.),width=321 ] the temperature dependence of the low - current and high - current differential resistances , @xmath58 and @xmath62 , are shown in fig . [ fig : rvst ] for two magnetic fields . between @xmath26 and a field dependent characteristic temperature @xmath61 , these two differential resistances are identical ( the @xmath33-@xmath34 curve is linear ) . at @xmath61 the two resistance curves bifurcate ; @xmath58 decreases rapidly with decreasing temperature and reaches zero ( @xmath63 @xmath64 ) at what is taken to be the melting temperature , @xmath65 , while the high - current resistance @xmath62 remains finite and varies smoothly across the melting line . the pinned liquid domain extends over @xmath66 where signs of pinning are still present . below @xmath65 , in the vortex solid phase , @xmath58 is zero and @xmath62 attains a finite , temperature - independent value visible in fig . [ fig : rvst ] . moreover , this value is the same for 1 t and 3 t and agrees reasonably with a linear extrapolation of the normal phase resistance to low temperature . more insight on the low - temperature saturation of the high - current resistance @xmath62 is gained from its field dependence for several temperatures shown in fig . [ fig : rvst ] . in the liquid phase , we find an approximately @xmath67 field dependence over the whole field range . in the solid phase , however , @xmath62 becomes independent of magnetic field above about 1 t. on one sample the measurements were extended to 17 t at 5 k and no variation of @xmath62 in excess of the experimental uncertainties of about 10 % were found in the field range 1 to 17 t , in contrast to the 17-fold increase predicted for the @xmath35 resistivity by eq . ( [ eq : bs ] ) . the differential resistance close to the threshold current , @xmath68 , behaves similarly ( see inset ) . in fact , in the parameter range where both quantities can be measured , we find a field and temperature independent proportionality between @xmath68 and @xmath62 . the behavior of the characteristic currents @xmath69 and @xmath70 further corroborates the similarities of the dissipation mechanisms close to threshold and in the high - current limit . the similarity is again underlined by the same @xmath71 field dependence for both quantities in the low temperature region . all these observations suggest that the same mechanism is responsible for the dissipation in the vicinity of the threshold current as in the high - current limit . the similarity of behavior is attributed to the way in which the resistive front propagates from the current contacts towards the middle of the sample as the current is increased@xcite . the current distribution reaches the resistive limit characterised by @xmath62 only after the two resistive fronts meet . as described earlier@xcite at low temperature there is a difference between field cooled ( fc ) and zero field cooled ( zfc ) preparation . the zfc data for both @xmath69 and @xmath70 exhibit a characteristic peak at the same line in the @xmath72 plane which defines a low temperature region where the fc prepared state is only metastable . both fc and zfc preparations give rise to the same @xmath62 . a first analysis to get our bearings consists of normalizing the resistance to the resistance measured in the normal phase and the field to @xmath73 a refinement on this is to estimate what the normal resistance would have been at the temperature of the measurement were the sample not superconducting . following this procedure and noting that when the magnetoresistance value saturates with field its value approximates the expected normal resistance , it has been hypothesised to be just that and the value used to better interpolate the normal resistance . for the upper critical field , we use the form @xmath74 $ ] yielding @xmath75 t / k @xcite . to estimate @xmath76 in the @xmath77 range , we use a quadratic interpolating function to connect with the @xmath78 range on the hypothesis that the field saturation value for @xmath62 corresponds to @xmath79 . for the very low field data below about @xmath80 t where we could not apply sufficient current to reach the @xmath62 asymptote , we used the temperature and field independent scaling factor between @xmath62 and @xmath55 found at higher fields to estimate @xmath62 from the measurement of @xmath81 the flux flow resistance data treated in this way are displayed in fig . [ fig : rperrn ] . ) as a function of the reduced magnetic field ( @xmath82 ) . the bs law for the flux flow resistivity is represented by a straight line and corresponds to the right hand scale.,width=321 ] it is evident from fig . [ fig : rperrn ] that the ratio of flux flow resistance to normal resistance is very different from the field proportional bs @xmath83 resistivity ratio of eq . ( [ eq : bs ] ) , ranging from a @xmath84 like variation at low field to a saturation value for the low temperature vortex solid phase . nonetheless there is a trend that with increasing temperature , in the vortex liquid phase , the saturation feature disappears , at least for the field values accessible to us . in the solid phase at low temperatures there is strictly no field dependence of @xmath62 whatsoever above about 1 t although as mentioned above its value is in good agreement with the linear extrapolation to @xmath85 of the normal resistance measured above @xmath26 . since the field where saturation occurs is in the order of @xmath86 , it might appear that for fields below saturation , the vortices move @xmath87 times faster than expected , or that the vortex friction coefficient is @xmath87 smaller , from a simple application of the bs relation . the experiment measures resistance and the ratio of resistances is only proportional to the ratio of resistivities if the current distributions are identical . this is not the case in anisotropic materials like bscco where the current penetration depth is determined by the anisotropy of the resistivity . the latter is not only different in the superconducting state but is also field dependent . from this point of view we might interpret the comparison made in fig . [ fig : rperrn ] as an indication that the current penetration is considerably less in the low temperature vortex solid phase , progressively approaching the normal phase penetration depth at higher temperatures in the vortex liquid phase . indeed in the liquid phase , the data are in rough agreement with bs law if the effects of inhomogeneous current distribution within the sample are taken into account . the voltage measured on the top plane of the sample is affected by the @xmath3-axis properties because they influence the distribution of the transport current within the sample @xcite . in the simplest approach we could suppose that well into the ohmic response regime the current is distributed as for a normal anisotropic ohmic conductor@xcite , a surmise which is borne out by numerical simulations based on the superconductor model of ref.@xcite . the penetration depth @xmath88 for a length @xmath89 between current injection points on the surface is then given by @xmath90 ( provided @xmath91 sample thickness ) and the resistance @xmath92 along the length of the sample will scale like @xmath93 . if @xmath94 of eq . ( [ eq : bs ] ) and the resistive response of the josephson coupled planes in the superconducting phase is represented by @xmath95 , one obtains @xmath96 . results close to @xmath26 are indeed well described by this form . putting @xmath97 the 75-k data in fig . [ fig : rperrn ] , for instance , are reasonably well fitted with @xmath98 , which , given the uncertainties of the simple resistive thick - sample estimate , we regard as rough agreement with bs law . the crux of the matter then , from this viewpoint , is in @xmath99 [ ptb ] . ( c ) shows the effect of spatial dephasing across the junction.,width=321 ] the @xmath3 axis conductivity @xmath100 can be obtained from the @xmath101 characteristics of the interplane josephson junctions . the usual representation of a josephson junction in a classical s - wave superconductor subject to a spatially uniform phase difference is illustrated on fig . [ josephson junction]a : the voltage response to current is nil up to the josephson current @xmath102 at which point the junction opens to @xmath103 and responds to further current according to the normal phase tunnel conductance @xmath104 which is related to the critical current by the ambegaokar - baratoff relation @xmath105 . for currents @xmath106 the @xmath3 axis resistivity is that apppropriate to the normal state @xmath107 assuming the penetration depth to be limited by the resistive anisotropy gives the result quoted above , but this regime would only be attained for extreme currents @xmath108 a@xmath109 corresponding to @xmath110 a for our samples . that is clearly not the regime which concerns us here . the answer to this problem seems to be that the josephson critical currents are much lower than the ambegoakar - baratoff prediction . measurements @xcite on micrometer size mesa stacks of bscco junctions at zero field show much lower critical currents , @xmath111 where the quasi - particle conductivity @xmath112 as @xmath113 at low temperature and @xmath114 mv yielding @xmath115 a@xmath116 as illustrated schematically on fig . [ josephson junction]b the opening of the junction is followed by a voltage response at higher currents with slope @xmath17 when @xmath117 but even this intermediate rgime is not observed : a penetration depth of @xmath118 nm corresponding to an anisotropy factor @xmath119 would involve @xmath0 planes and a voltage response of @xmath120 v , still a factor of @xmath87 higher than observed in our measurements . to understand what we see , we must take account of the non - uniformity of the phase across the planes due to the random positioning of vortices from plane to plane . the critical current for the junction is then reduced to @xmath121 @xmath122 where @xmath123 represents the phase difference across the junction at position @xmath124 in the plane and @xmath125 the value of @xmath126 averaged over the junction area on a josephson penetration length scale@xcite . it is typically of order @xmath127 as illustrated on fig . [ josephson junction]c , the junction will then open to a potential difference of @xmath128 due to phase slippage at @xmath129 , beyond which it should continue to see a tunneling conductance @xmath130 of quasi - particles in the gap until @xmath131 when the tunneling of quasi particles at the gap edge will result in the usual normal phase dynamic conductance . the gap quasi - particle conductance is thus experienced over a wide current range between @xmath132 and @xmath133 over which the resistance ratio is considerably enhanced to @xmath134 where @xmath135 at low temperature@xcite , again if we assume eq . ( [ eq : bs ] ) to be valid for the second square root term . the magnetic field dependence of @xmath136 was measured in a subsequent experiment@xcite , also on micrometer size mesa samples . if finally we include also the effect of quasi - particle conductance @xmath137 in the @xmath35 plane , we find @xmath138 this is the expression that we compare with the experimental results on figs . [ comparison](a)-(d ) . equation ( [ eq : rfperrn ] ) is compared with the experimental findings for both the optimal doped and the underdoped samples on figs . [ comparison ] . for simplicity a single plot of eq . ( [ eq : rfperrn ] ) is drawn as a dashed line on the log - log plot against @xmath139 where @xmath140 is the temperature corrected value as described above . it corresponds to a choice of parameter values appropriate to the optimally doped situation at low temperature @xmath141 : @xmath142 t , @xmath143@xcite , @xmath20 , @xmath144@xcite , @xmath145 and @xmath146 corresponding to a linear extrapolation to @xmath85 . a parallel solid line is also drawn as the best @xmath147 fit to the data . we note that : \1 . the order of magnitude is correct to within a factor of 2 for both optimal and underdoped samples , with the notable exception of the saturation plateau values in the vortex solid phase , especially in the underdoped sample . the square root behavior predicted by the bs relation is approximately correct , again with the obvious exception of the plateaus . the magneto - resistance plateaus are all situated in the vortex solid phase . it is to be remarked that the @xmath148 k magneto - resistance in the underdoped sample begins to saturate , but on reaching about @xmath80 t it reverts to an increasing power law . the melting field for @xmath148 k is estimated to be about @xmath149 t from the appearance of finite @xmath58 . clearly equation ( [ eq : rfperrn ] ) does not describe all the features of the magneto - resistance , even if it does reasonably well for temperatures above the vortex solid melting and in the solid region for fields below the saturation . it has a structure which could in principle describe a plateau if the @xmath150 coefficient in the field dependence of the shunt conductance of the josephson coupled layers were to be much larger @xmath151 such as might be introduced by aligned vortex core nin tunneling junctions or possibly @xmath88-wave node quasi - particle to vortex core tunneling . such effects have not however been seen in the mesa magneto - conductivity measurements@xcite , although these were performed at relatively high temperature ( @xmath152 k ) and high field , so in the vortex liquid phase . it is difficult to conclude that the bs relation is untrue , either by a multiplicative factor or in its form . to determine the multiplicative factor requires a good way of estimating the normal resistivities in the superconducting phase also , because the agreement is reasonably good at high temperature and the vortex entity is not expected to change nature from high to low temperature , the serious disagreement at low temperature is much more likely to be related to the nature of the solid phase as the bs relation is a single vortex property . it can however be concluded from the sublinear , approximately @xmath67 , field dependence that if the linear field dependence of the bs relation is correct , the current distribution is not constant and its penetration depth is controlled by the resistive anisotropy and thus by @xmath16 . we acknowledge with pleasure fruitful discussions with f.portier , i.tutto , l.forro and t.feher and the help and technical expertise of f.toth . l.forro and the epfl laboratory in lausanne have contributed in a very essential way to sample preparation and characterisation . finally we acknowledge with gratitude the otka funding agency for otka grant no .	measurement of the free flux flow resistance in monocrystalline highly anisotropic bscco reveals a low field magneto - resistive effect nearly @xmath0 times that expected from a naive application of the bardeen - stephen rule @xmath1 . furthermore in the vortex solid phase it saturates to a constant value at about @xmath2 as if the vortices were moving @xmath0 times faster than expected . an attempt to account for this behavior by a combination of sheared planar vortex flow obeying the bs rule and the quasi - particle @xmath3 axis conductivity between the josephson coupled superconducting planes is only partially successful . in particular it is unable to account for the saturation feature which occurs in both optimally and underdoped samples in the low temperature vortex solid phase .
the broad - line region ( blr ) is one of the main components of an unobscured agn ( ie , those where we have an unimpaired view torwards the nucleus ) , and is charactized by very broad emission lines in the uv / optical spectra ( figure 1 ) with typical widths of several thounsands km / sec due to the proximity to the central bh . one of the features of this region is that the observable uv / optical continuum flux variations are closely related to variations of the emission - lines fluxes . combining the correlation between continuum and line fluxes variations with the widths of the lines , the mass of the central black hole can be estimated through the reverberation mapping technique ( see 2 ) . + in these days , when it is widely accepted that all massive galaxies harbor a black hole in their centers , reverberation has allowed real physical comparison between active and dormants bhs in the local universe , but so far has not probed the high end of the agn luminosity range . + in the next section we describe briefly the basic theory and results from previous reverberation mapping studies leading to the motivation for a high - luminosity high - redshift quasars monitoring campaign that we are performing since the beggining of 2005 . our main aim is to extend these results by two orders of magnitude , giving the firsts measurement of the largest bhs , and thus extending our knowledge of the physics of agn and their hosts into an epoch crucial for the understanding of galaxy evolution . reverberation mapping is a technique based on the response of the blr gas to changes in the central continuum source . it is based on three basic assumptions : 1 . _ the ionizing continuum originates in a single central source_. + this holds because the blr size is @xmath02 orders of magnitude larger than the continuum source . the light travel time across the blr is the most important time scale . _ + which means that the clouds response is practically `` instantaneous '' . indeed , the recombination time is related with the electron density @xmath1 by the relation @xmath2 , which is much less than the typical light travel time across the blr . there is a simple but not necessary linear relationship between the observed continuum and the ionizing continuum ( which gives rise to the emission lines ) . _ from these points one can construct a simple 1d linear model given by @xmath3 where @xmath4 and @xmath5 are the line and continuum light curves respectively . @xmath6 is the _ transfer function _ ( tf ) which holds the information about the geometry and kinematics of the blr gas . due to technical limitations , the tf is collapsed to one single parameter : the time lag @xmath7 between both continuum and line light curve variations which is obtained through a cross - correlation analysis between them ( peterson 2001 ) . this lag @xmath7 is used to directly measure the blr size from @xmath8 ( peterson et al . , 2004 , kaspi et al . , 2005 ) . assuming a gravitationally bound system ( proven to be correct for those objects with measurements from several lines , where the observed anticorrelation between the distance @xmath9 to the bh and the line doppler widths is consistent with virialized motions of the blr gas ) and measuring the blr line widths @xmath10 , it is possible to infer the mass of the central bh through @xmath11 where @xmath12 depends on the geometry of the blr and it is expected to be @xmath01 . many important results have been found from reverberation mapping studies , such as the stratification of the blr and black hole masses . also , an empirical relation between the blr size and the agn luminosity has been determined , which shows that _ the size of the broad - line region scales with luminosity_. this is not far from the @xmath13 relation expected if the blr in all agn have similar @xmath1 value and ionization conditions . kaspi et al . ( 2005 ) found that the scaling relation has the form @xmath14^{0.6\pm0.1}\mbox { pc}\ ] ] ( as it is shown in figure 1b ) . preliminary results by bentz et al ( 2006 ) suggest a smaller slope of about 0.52 . using @xmath15 relation is thus possible to estimate the blr radius using @xmath16 alone , hence making bh mass estimation accessible for a large number of agn . such results can not be directly applied to high-@xmath17 , high-@xmath16 sources that contain the most massive bhs , since measuring their blr size requires an extrapolation by up to two orders of magnitude in luminosity in eq . ( [ rl ] ) , which would be subject of a large uncertainty in the measured @xmath18 ( and then in the bh mass ) . this is the starting point of our quasars monitoring program . the few attempts to extend the luminosity range of agn with reverberation mapping results to high - luminosity agns have failed for two reasons : 1 . most high-@xmath16 sources are very slow variables with very low amplitude ( < 20% ) . 2 . the line luminosity amplitude is even smaller than the continuum amplitude because the emission line response is averaged over the very large ( @xmath19 pc ) blr geometry . since 2005 we are undertaking a novel strategy that uses optical broad band imaging to trigger spectroscopic follow up on a time scale of about one year . the idea is to identify high-@xmath16 sources with large enough continuum variations that will enable accurate and meaningful measurement of emission line variations . only a small fraction of high-@xmath16 sources show such variations at any given time , hence , a large number of sources must be constantly monitored to identify the onset of such events . given large enough continuum variations of a certain agn , the event must be followed spectroscopically . the expected blr size suggests that the spectroscopy phase should continue for at least four years . our sample consists of 56 quasars which are the brighest from the calan - tololo , 2df and sdss catalogues spread over all ras , with @xmath20 erg / sec , redshifts that range from 2.3 to 3.3 and mean r magnitude about @xmath018.2 . until today , black hole mass estimations using reverberation mapping analysis have been mainly based in the variability of the h@xmath21 ( @xmath224861 ) emission line which lies in the optical . in our optical monitoring program we are looking for the civ ( @xmath221549 ) emission line variability which , because of the the high redshifts of the ojects , falls into the optical window . reverberation mapping based in the h@xmath21 line can be made for these objects going to the ir . imaging is obtained with the 0.9 m telescope at cerro tololo interamerican observatory ( ctio ) every month for a sub - sample of quasars to monitor their continuum emission in the r band . monthly datasets allow us to have from 4 to 5 light curves points in average per object in a year , enough to detect the onset of a high variation in their fluxes . + we use differential photometry respect 8 to 13 field stars which allows a photometry independent of the weather conditions , since the object and the field stars are observed simultaneously . these field stars are selected near to the quasar in order to appear in all the observations and have a magnitude cut - off of order of the quasar magnitude . to apply differential photometry we average the field star magnitudes in order to minimize the random variability of each of them , then @xmath23 is the average field star magnitude . the _ rms _ of each field star light curve is also checked periodically to look for intrinsically variable objects . this procedure gives us an average magnitude almost constant within the error bars ( @xmath24 variable with a typical value of @xmath25 ) . the quasar differential light curve ( dlc ) is given by @xmath26{\pi_{i}f_{ij}}}\right)\ ] ] where @xmath27 is the difference in magnitude between the quasar and the average of field stars at time @xmath28 ( in flux it reads as the ratio of the object flux with the geometrical average of the field stars ) . so , as @xmath29 is almost flat , the time series of @xmath27 gives us the quasar light curve with respect to a constant standard and then variations of @xmath30 are exclusively atributted to the quasar . to quantify variability we define the paremeter @xmath31 where @xmath32 and @xmath33 are the maximum and minimum flux in the dlc , then @xmath34 is a function of the difference @xmath35 ( because @xmath36 ) . in this way , the @xmath34 parameter is just the peak - to - peak variation in flux . + a minimum variability is needed in the continuum to detect a correlated variability in the emission lines ( which variation is weaker because is averaged in a larger geometry than the continuum source ) . previous reverberation campaigns have concluded that this minimum is @xmath015% , then , objects with variability larger than that trigger the spectroscopic follow - up . as mentioned before , we are triggering the spectroscopy follow up whenever we detect 15% variability using the 2.5 m dupont telescope at las campanas obseratory ( lco ) . we note that at such quasar luminosity , the typical intrinsic continuum variability time scale is 510 months . given the @xmath37 factor , in the observed frame the lines are expected to follow the continuum after 12 years . during 2005 - 2006 we obtained spectroscopic observations for the whole sample to establish a zero baseline for future spectroscopy and evaluate the spectroscopic accuracy of our method . in 2007 we have started the actual spectroscopic follow - up of the first quasars . currently , there are 5 targets which have been followed up since 2007 and there are 3 more quasars reciently ( 2008 ) selected for the spectroscopic monitoring . after 3 years of monitoring we can confirm that , in general , quasars are very slow and low amplitude variables ( < 0.1 mag ) . however , our strategy has allowed us to identify a subsample of 8 quasars that present a variability greater than 15% ( about 0.2 magnitudes , shown in figure 2 ) which correspond to the 14.2% of the sample . the variability in some of the objects has been seen to reach up to 20% and spectroscopic follow up has began for them . on the other hand , the varibility of the field stars @xmath38 is lower than 3% . this range is represented in figure 2 using dot - dashed lines . all points of @xmath38 fall within this region . + up to date dlc for our targets can be found at the website of the program ` http://www.cec.uchile.cl/\simibotti/reverberation/ ` . we reviewed the basis of reverberation mapping technique which is a powerful tool to obtain black hole masses in agn and the size - luminosity ( @xmath39 ) relation from which we can estimate the size of the blr ( and then estimate the bh mass ) from a single epoch observation . currently this relation covers almost 5 magnitudes in luminosity , but it is subjected to large uncertainties when is extrapolated to high luminosity agn . for a better determination of the slope of this relation we need to span the luminosity range of agn with reverberation mapping results of @xmath40 erg / sec . to achieve that aim we present a monitoring campaing of a large sample of high-@xmath17 quasars which will extend these results by two orders of magnitude , measuring the broad - line region size and black hole mass of luminous agn at high redshifts . our strategy has been successful identifying a subsample of high variability ( @xmath4115% ) quasars which are being followed spectroscopically since 2007 . + we expect that in the next 3 - 6 years reverberation mapping analysis will finally yield the highly anticipated measurement of bh masses in some of the most luminous quasars in our universe . + we acknowledge the financial support from the fondecyt grant no . 1080603 and proyecto fondap de astrofsica 15010003 . peterson , b. , 2001 , in `` the starburst - agn connection '' , ed . by i. arextaga , d. kunth , r. mjica . peterson , b. et al . , 2004 , apj , * 613 * , p682 . kaspi , s. et al . , 2005 , apj , * 629 * , p61 . bentz , m. et al . , 2006 , apj , * 644 * p133 .	in this contribution we briefly review the reverberation mapping technique and its results for low and intermediate luminosity agns . then we present a monitoring campaign of high - luminosity high - redshift quasars which will extend these results by two orders of magnitude , probing the broad - line region size and black hole ( bh ) mass of luminous agn at redshift @xmath02 - 3 . 1.0 cm
define a knot or link in @xmath0 to be hyperbolic if its complement is a hyperbolic @xmath1-manifold . this implies that there is a covering map @xmath2 from @xmath3 to @xmath4 such that the covering translations are isometries of @xmath3 . we say that an embedded or immersed surface @xmath5 in @xmath4 is totally geodesic if it is isotopic to a surface that lifts to a set of geodesic planes in @xmath3 . embedded totally geodesic surfaces are relatively rare in knot and link complements . it was conjectured in @xcite that there are no closed embedded totally geodesic surfaces in hyperbolic knot complements . this has been verified for hyperbolic knots that are alternating knots , tunnel number one knots , and 2-generator knots @xcite , almost alternating knots @xcite , toroidally alternating knots @xcite , montesinos knots @xcite , 3-bridge knots and double torus knots @xcite ) and knots of braid index three @xcite and four @xcite . note that the conjecture does not hold for links . the first explicit counterexample was given in @xcite . in @xcite , it is demonstrated that for each @xmath6 , there exists a 2-component hyperbolic link that contains a closed embedded totally geodesic genus @xmath7 surface . in this paper , we investigate surfaces that are not closed in knot and link complements . in particular , we are interested in totally geodesic seifert surfaces , where a seifert surface for a knot or link @xmath8 is a compact orientable surface whose boundary is @xmath8 . note that although a totally geodesic seifert surface need not necessarily be minimal genus , it is incompressible . work of thurston implies that a properly embedded surface @xmath5 with boundary in a hyperbolic knot exterior @xmath9 can have one of three possible behaviors : 1 . there is a nontrivial simple closed curve @xmath10 in @xmath5 that is not boundary parallel in @xmath5 but that corresponds to a parabolic isometry . this implies that @xmath10 is isotopic into a neighborhood of the missing knot . in this case , we say that @xmath10 is an accidental parabolic curve and @xmath5 is an accidental surface . see @xcite for more on these surfaces . @xmath5 is a quasi - fuchsian surface , with limit set a union of quasi - circles . this includes the case of @xmath5 being totally geodesic ( also called fuchsian ) , in which case the limit set is a union of geometric circles @xmath11 is a fibered knot and in some finite cover , @xmath5 lifts to a fiber . in this case , the limit set is the entire sphere . in @xcite , it is proved that if @xmath11 is not fibered , then any minimal genus seifert surface can not be accidental and hence must be quasi - fuchsian . see also @xcite.a quasi - fuchsian surface @xmath5 can admit an essential annuli @xmath12 intersecting @xmath5 in its boundary . however , in this case , @xmath5 can not be totally geodesic . in essence , the existence of the annulus forces pairs of quasi - circles in the limit set of the cover to share two points , which a pair of circles that do not cross can not do . so , for a surface to be totally geodesic , it must not admit any essential annuli . nor can it admit an essential disk with boundary consisting of four arcs , two of which lie in the surface and two of which lie in the cusp . such a disk doubles to an essential annulus in the manifold obtained by doubling along the totally geodesic surface . however this contradicts the fact the double is a hyperbolic manifold . a seifert surface is said to be _ free _ if the complement of a neighborhood of the seifert surface is a handlebody . note that there are examples of knots such that a minimal genus seifert surface is free and examples such that it is not free . see @xcite . in section [ s : tgss ] , we explain how lifts of rigid hyperbolic 2-orbifolds can be utilized to generate totally geodesic seifert surfaces . in section [ s : examples ] , we utilize these results to generate the first examples of hyperbolic knots with totally geodesic seifert surfaces . we include both examples of free totally geodesic seifert surfaces and non - free totally geodesic seifert surfaces . in section [ s : twobridge ] , we prove that many hyperbolic knots can not possess a totally geodesic seifert surface , including all two - bridge knots . note that the surfaces given here are the only known examples of embedded totally geodesic surfaces of any kind in knot complements in @xmath0 . one can ask whether there are embedded totally geodesic surfaces in knot complements other than seifert surfaces . acknowledgements : thanks to satyan devadoss , cameron gordon and alan reid for helpful conversations . we call a hyperbolic orbifold _ rigid _ if there exists only one possible hyperbolic metric that generates the orbifold . in @xcite , it is demonstrated that a hyperbolic 2-orbifold @xmath13 with underlying space @xmath14 , @xmath15 elliptic points and @xmath16 corner reflectors is rigid if and only if @xmath17 . a 2-orbifold is hyperbolic if and only if its euler characteristic is negative where the euler characteristic is given by : @xmath18 where @xmath19 and @xmath20 are the orders of the elliptic points and corner reflectors respectively . hence a complete list of rigid hyperbolic 2-orbifolds consists of * @xmath21 for @xmath22 , * @xmath23 for @xmath24 * @xmath25 for @xmath26 note that each @xmath27 and @xmath28 can be either a positive integer or @xmath29 . the fundamental group @xmath30 of an orbifold @xmath31 is the group of deck transformations of the universal cover @xmath32 . a @xmath33-orbifold @xmath13 embedded in a @xmath1-orbifold @xmath31 is _ essential _ if @xmath34 injects into @xmath30 . a hyperbolic 2-orbifold @xmath13 embedded in a hyperbolic 3-orbifold @xmath31 is _ totally geodesic _ if it lifts to the disjoint union of geodesic planes in the universal cover @xmath3 . let @xmath2 denote the map projecting @xmath3 onto @xmath31 . we state a theorem that follows from the theorem on p. 217 @xcite , but provide a complete proof here for convenience . [ main ] let @xmath13 be an essential rigid hyperbolic @xmath33-orbifold embedded in a hyperbolic @xmath1-orbifold @xmath31 . then @xmath13 is isotopic relative the singular set to a totally geodesic @xmath33-orbifold @xmath35 . every isometry @xmath36 of hyperbolic 3-space can be represented as the product of two order two elliptic isometries , each a rotation about a geodesic axis . if @xmath36 is hyperbolic , then the two axes , considered at subsets of @xmath37 , do not intersect . if @xmath36 is parabolic then they share an endpoint . if @xmath36 is elliptic , then they intersect in @xmath3 and the axis of rotation for the elliptic isometry passes through the point of intersection and is perpendicular to the plane containing both of the axes . the fundamental group @xmath30 is realized as a discrete group @xmath38 of isometries of @xmath3 with subgroup @xmath39 corresponding to @xmath34 . there exists a pair of isometries @xmath40 and @xmath41 in @xmath39 , the product of which is a third isometry @xmath42 in @xmath39 . in the cases of @xmath21 and @xmath23 , we can choose these isometries corresponding to the three singularities and the three corner reflectors respectively . in the case of @xmath25 , we can choose isometries corresponding to the corner reflector and to two lifts of the isometry corresponding to the singularity , one of which is conjgate to the other by an order two elliptic isometry corresponding to the boundary of the orbifold . each @xmath43 can be realized as the product of two isometries , each of which is an order two elliptic isometry about an axis . choose an axis for an order two elliptic isometry @xmath44 such that if @xmath40 is elliptic , the axis is perpendicular to the axis of @xmath40 . if @xmath40 is parabolic , choose the axis of @xmath44 so that it ends at the fixed point of @xmath40 . also make the axis simultaneously satisfy the same requirements for @xmath41 . then we can find order two elliptic isometries @xmath45 and @xmath46 such that @xmath47 and @xmath48 . note then that @xmath49 . since @xmath42 is elliptic or parabolic , the axes for @xmath45 and @xmath46 must either intersect in @xmath3 or in @xmath50 . in either case , the three axes for @xmath45 , @xmath44 and @xmath46 all lie in a geodesic plane @xmath51 . the entire group @xmath39 is generated by @xmath45 , @xmath44 and @xmath46 so @xmath39 preserves @xmath51 . the orders of @xmath40 , @xmath41 and @xmath42 determine the angles of the triangle defined by the axes of @xmath45 , @xmath44 and @xmath46 , and hence the resulting 2-dimensional hyperbolic orbifold @xmath35 that is the projection of @xmath51 to @xmath31 is rigid . exactly as in the proof of theorem 3.1 of @xcite , we can prove that the limit set of @xmath51 is in fact also the limit set of a topological plane @xmath52 covering @xmath13 . since @xmath13 is embedded , the limit sets of the topological planes covering @xmath13 can only intersect tangentially . hence the same holds for all the copies of @xmath51 under the group action , implying that the geodesic planes covering @xmath35 do not intersect in their interiors . in particular , this means that @xmath35 is embedded and @xmath53 is a disjoint union of geodesic planes , and @xmath35 is a totally geodesic orbifold . suppose @xmath54 is an element of @xmath38 such that @xmath55 . then @xmath54 preserves the limit circle of @xmath51 . but then @xmath54 preserves the limit set of @xmath52 , implying @xmath56 . hence , @xmath13 and @xmath35 have the same fundamental group and are orbifold homeomorphic . note that in the case the underlying space of @xmath35 is a disk , @xmath13 and @xmath35 must share the boundary arcs , which are order two singularity axes in @xmath31 . the orbifold @xmath13 can be isotoped so that it does not intersect @xmath35 in its interior as follows . first , put @xmath13 in general position relative to @xmath35 . utilizing the product neighborhoods of the singularities , cusps and boundaries of @xmath35 , we can isotope @xmath13 so that there are no intersections curves near singularities , cusps , or boundaries of the two 2-orbifolds . hence , all the intersection curves are simple closed curves . a trivial intersection curve on one of the 2-orbifolds must be trivial on the other by the fact @xmath13 is essential and @xmath35 is totally geodesic . however , such a curve can then be eliminated by isotoping @xmath13 . a curve that is nontrivial on both @xmath13 and @xmath35 , and innermost such on @xmath13 must bound a disk with one singularity on each of @xmath13 and @xmath35 , denoted @xmath57 and @xmath58 respectively . since this implies that a loop around the singularity on @xmath13 is homotopic to the loop around the singularity on @xmath35 , the singularities must either both correspond to parabolic isometries or to elliptic isometries of the same order . then @xmath59 forms a sphere with two singularities of the same order . by irreducibility , @xmath59 bounds a ball containing an unknotted singular arc or cusp within it . hence , @xmath57 can be isotoped through @xmath58 to eliminate the intersection . now , @xmath13 and @xmath35 do not intersect . hence their lifts , which come paired with shared boundary circles , do not intersect . thus each pair of planes sharing a boundary circle bound a ball in @xmath3 that is intersected by none of its translates in the interior . thus , we can isotope the topological plane that is a lift of @xmath13 to the corresponding geodesic plane that is a lift of @xmath35 . in fact , we can do so equivariantly with respect to @xmath38 , and none of the disjoint balls bounded by the planes will intersect in the process . hence @xmath13 is isotopic to @xmath35 , as we wished to demonstrate . [ c : totallygeod ] let @xmath9 be a closed orientable 3-manifold , covering a 3-orbifold @xmath31 such that @xmath60 contains a link @xmath8 covering a collection @xmath61 of simple closed curves and arcs beginning and ending on the order two singular set of @xmath31 , and such that @xmath62 is a hyperbolic 3-orbifold @xmath63 . then if @xmath63 contains a rigid hyperbolic 2-orbifold @xmath13 for which each simple closed component of @xmath61 is a single puncture ( singularity with @xmath64 ) and each arc corresponds to a corner reflector with @xmath65 , then @xmath13 is isotopic in @xmath63to a 2-orbifold that lifts to a totally geodesic seifert surface for the link @xmath8 in @xmath9 . this follows immediately from theorem [ main ] . in this section , we utilize corollary [ c : totallygeod ] to give a variety of examples of totally geodesic seifert surfaces in knot and link complements in @xmath0 . balanced pretzel links the seifert surfaces for this family of balanced pretzel links are all totally geodesic.,height=192 ] let @xmath9 be @xmath0 and @xmath31 the orbifold obtained by quotienting out by a @xmath66 rotation about a great circle @xmath67 in @xmath0 , where @xmath68 . let @xmath57 be a disk with two singularities of order @xmath2 and @xmath2 , twisted with @xmath69 crossings as in figure [ fig : balancedpretzel ] . let @xmath61 be the boundary of the disk @xmath57 . then @xmath61 lifts to an @xmath70 pretzel knot @xmath11 , where @xmath69 appears @xmath2 times . we call @xmath11 a _ balanced pretzel knot_. a picture of the horoball diagram for the @xmath71 pretzel knot appears in figure [ fig : balancedpretzellift ] . this picture was generated by jeffrey weeks snappea computer program , ( cf . the sequentially tangent horoballs in a line correspond to a vertical geodesic plane coming up out of the xy - plane that covers the totally geodesic surface . the straight line of horoballs corresponds to the totally geodesic surface.,height=4 ] polyhedral links a link that projects to the surface of the cube , as in figure [ fig : cubelink ] , corresponds to the loop in the orbifold shown , which possesses the shaded rigid 2-orbifold that appears in its complement . therefore , the link bounds the totally geodesic seifert surface shown . a totally geodesic seifert surface for this cube link.,height=2 ] knots with totally knotted seifert surfaces instead of a disk with two cone points that is twisted relative the singular arc as we saw in figure [ fig : balancedpretzel ] , we will choose a disk that is knotted relative the singular curve as in figure [ fig : totallyknotted ] . when @xmath72 , this lifts to the knot in figure [ fig : totallyknotted ] . the disk lifts to a totally geodesic seifert surface , which is totally knotted , since the complement of a regular neighborhood of the surface is thurston s wye manifold , which is known to be hyperbolic with incompressible boundary . a totally knotted seifert surface which is totally geodesic.,height=2 ] we prove the following theorem . note that the only non - hyperbolic @xmath33-bridge knots are the @xmath73-torus knots . there are no non - boundary parallel incompressible closed surfaces in their complements . in @xcite , a complete classification of all the incompressible boundary - incompressible surfaces in the complements of two - bridge knots is given . all such surfaces are carried by branched surfaces that appear as in figure [ fig : branchedtwobridge ] . such a surface is orientable if and only if each @xmath74 is even where @xmath75 $ ] . for each @xmath76 , there is a unique choice of such even @xmath74 s . note that a seifert surface corresponds to a @xmath77-sheeted surface carried by the branched surface . each possible seifert surface is obtained as a collection of horizontal `` plumbing '' disks either to the inside or the outside and vertical strips between them , each with an even number of half - twists . the other incompressible boundary - incompressible surfaces have more than one sheet . if for some @xmath78 , we have @xmath79 , then there is a disk in @xmath0 with boundary that consists of two arcs on @xmath5 and two arcs on the cusp boundary . when the manifold is doubled along @xmath5 , this disk doubles to an essential annulus in the resulting manifold . hence , the resulting manifold can not be hyperbolic , implying that the surface @xmath5 could not be totally geodesic . when @xmath80 for all i , choose a curve @xmath10 on @xmath5 as in figure [ fig : branchedtwobridge ] . note that the curve @xmath10 can be isotoped to @xmath81 as in figure [ fig : branchedtwobridge ] . the curve @xmath81 sits on a torus @xmath82 in @xmath83 . cutting the torus open along @xmath81 yields an annulus in @xmath83 , both boundary components of which lie on @xmath84 in @xmath82 . gluing on to its boundary components a pair of parallel annuli that give the isotopy from @xmath10 to @xmath81 yields an annulus @xmath12 in @xmath85 with boundary in @xmath86 . the annulus @xmath12 is incompressible in @xmath87 since its core curve is a power of a generator for the fundamental group of the complementary handlebody . to see that it is boundary - incompressible , we can choose the essential arc @xmath88 in @xmath12 shown in figure [ fig : essentialarc ] . if @xmath88 can be isotoped into the surface , then there exists a disk with boundary consisting of two arcs on @xmath5 and two arcs on the cusp boundary . again , when the manifold is doubled along @xmath5 , this yields an essential annulus , a contradiction to the hyperbolicity of a hyperbolic manifold doubled along a totally geodesic surface within it . if @xmath88 can not be isotoped into the surface , then @xmath12 is an essential annulus intersecting @xmath5 in its boundary , which obstructs @xmath5 from being totally geodesic . do two - bridge knots possess non - orientable totally geodesic surfaces ? any essential non - orientable surface for a 2-bridge knot or link is given by a choice of @xmath89 $ ] , where @xmath75 $ ] and at least one @xmath74 is odd . the only candidates for totally geodesic surfaces from amongst these are those where each @xmath74 is odd . in the case of links , these surfaces can be totally geodesic . the whitehead link possesses such a surface . however , we have yet to see such a surface in a 2-bridge knot complement that is totally geodesic . menasco , william and reid , alan , totally geodesic surfaces in hyperbolic link complements " , _ topology 90 ( columbus , oh , 1990 ) _ , 215 - 226 , ohio state univ . publ . , 1 , de gruyter , berlin , 1992 .	the first examples of totally geodesic seifert surfaces are constructed for hyperbolic knots and links , including both free and totally knotted surfaces . then it is proved that two bridge knot complements can not contain totally geodesic orientable surfaces .
the discovery and ongoing development of gauge / string dualities @xcite has opened up a new and exciting frontier for nonperturbative qcd . already the current , first generation of `` holographic qcd '' or `` ads / qcd '' duals with their bold approximations is beginning to provide new and often surprising analytical insights into the elusive infrared ( ir ) sector of the strong interactions . by now a rather large set of static hadron properties has been calculated in the ads / qcd framework ( for recent reviews see e.g. ref . [ 2 ] ) . the majority of this work was based on the two currently most popular dual candidates , i.e. the hard- @xcite and soft - wall @xcite backgrounds . since even these rather minimal gravity backgrounds turn out to describe most calculated static hadron properties at an astonishing 10 - 30% accuracy level , it becomes increasingly important to explore their capacity and limitations in describing more detailed and sensitive qcd amplitudes . one such class of amplitudes comprises the hadron form - factors , and several of them have already been estimated holographically @xcite . another important set of hadron amplitudes are the @xmath0-point functions of hadronic interpolating fields , and among these the two - point correlators play a special role , not least because detailed qcd results are available for most of them . we have therefore recently advocated @xcite to put ads / qcd dual candidates to more stringent tests by evaluating their predictions for hadron correlators and by confronting those with qcd information from the lattice @xcite , the operator product expansion ( ope ) including hard instanton contributions to the wilson coefficients @xcite , a hypothetical uv gluon mass suggested to encode the short - distance behavior of the static quark - antiquark potential @xcite , and a scaling low - energy theorem @xcite based on the trace anomaly . ( ads / qcd correlators were recently also studied in refs . [ 1113 ] . ) in the following we will outline the main steps of implementing this program in the scalar glueball channel @xcite . to this end , we review the calculation of the scalar glueball spectra , decay constants and correlators in both hard- and soft - wall backgrounds , and we comment on a recent attempt @xcite to generalize the soft - wall correlator . the holographic hard- and dilaton soft - wall duals are both based on five - dimensional bulk geometries of `` poincar domain wall '' type@xmath1where @xmath2 is the four - dimensional minkowski metric and conformal invariance of the dual gauge theory in the uv requires @xmath3 . the soft wall additionally contains a bulk dilaton field @xmath4 . scalar qcd glueballs are interpolated by the lowest - dimensional gluonic operator carrying vacuum quantum numbers , @xmath5(where @xmath6 is the gluon field strength ) . since @xmath7 has conformal dimension @xmath8 ( at the classical level ) , the ads / cft dictionary @xcite prescribes its dual string modes @xmath9 to be the normalizable solutions of the scalar wave equation in the bulk geometry ( [ metric ] ) ( and possibly other background fields ) with the uv behavior @xmath10 . the latter implies that the square mass @xmath11 of the bulk field @xmath12 vanishes for @xmath13 , so that its minimal action takes the form@xmath14 = \frac{1}{2\kappa ^{2}}\int d^{5}x\sqrt{% \left\| g\right\| } e^{-\phi } g^{mn}\partial _ { m}\varphi \partial _ { n}\varphi . \label{sd}\ ] ] the four - dimensional fourier transform @xmath15 of the normalizable dual modes solves the reduced field equation@xmath16 \hat{\varphi}\left ( q , z\right ) = 0 \label{fe}\]]obtained by variation of the bulk action ( [ sd ] ) . the corresponding orthonormalized solutions @xmath17 have discrete momenta @xmath18 which determine the glueball mass spectrum of the boundary gauge theory . for dilaton fields which vanish at the uv boundary ( as in the soft wall case ) , the glueball correlator has the spectral representation @xmath19where a prime denotes differentiation with respect to @xmath20 and regularizing contact terms for @xmath21 are not written explicitly . the pole residues of eq . ( [ specrep ] ) contain the decay constants @xmath22of the @xmath0-th @xmath23 glueball excitation . since the @xmath24 can be regarded as the glueball ( bethe - salpeter ) wave functions at the origin , a smaller glueball size implies a higher concentration of the wave function and consequently a larger value of @xmath24 . evidence for such an enhancement of the ground - state decay constant was found in instanton vacuum models sch95 , in qcd sum rule analyses which include instanton contributions to the ope coefficients @xcite , and in ( quenched ) lattice simulations @xcite . the hard - wall geometry of polchinski and strassler @xcite is an ads@xmath25 slice with a randall - sundrum type cutoff @xmath26 at the ir brane , @xmath27which implements conformal symmetry in the uv and its breaking in the ir in a minimal fashion . the glueball decay constants in the hard - wall background can be calculated directly from the normalized solutions@xmath28(where @xmath29 ) of the field equation ( [ fe ] ) in the metric ( [ hw ] ) . the constants @xmath30 are determined by the normalization condition @xmath31 . for the dirichlet ir boundary condition @xmath32 one then obtains @xcite @xmath33while the alternative neumann boundary condition @xmath34 yields @xcite @xmath35here @xmath36 denotes the @xmath0-th zero of the @xmath37-th bessel function abr72 . from the general expression ( [ dc ] ) for the decay constants and the hard - wall eigenmodes ( [ hwm ] ) one then finds @xcite@xmath38or more specifically for the above two ir boundary conditions@xmath39for a quantitative estimate one can fix the overall normalization factor @xmath40 according to eq . ( [ est ] ) and set the ir scale @xmath41 e.g. such that a typical quenched ground - state glueball mass of @xmath42 1.5 gev is reproduced , or at the value @xmath43 gev found in the classical hadron sector . either way , the ground - state decay constant predictions remain in the range @xmath44 gev @xcite . the hard - wall predictions ( [ mnd ] ) , ( [ mnn ] ) for the squared masses of scalar glueballs ( and of other hadrons ) grow quadratically with high radial ( and orbital ) excitation quantum numbers , in contrast to the linear trajectories expected from semiclassical models and data . the dilaton soft - wall background @xcite@xmath45provides an economical corrective to this problem in the meson @xcite and glueball @xcite sectors . ( the `` metric soft wall '' @xcite is a dilaton - less alternative which also yields linear baryon mass trajectories . ) the spectrum - generating normalizable solutions of eq . ( [ fe ] ) in the background ( [ sw ] ) are those kummer functions whose power series expansion truncates to generalized laguerre polynomials @xmath46 @xcite , i.e.@xmath47where @xmath48 , @xmath49 and @xmath50 is a confluent hypergeometric function @xcite . the ensuing restriction to discrete @xmath18 generates the mass gap @xmath51 and the glueball mass spectrum @xcite@xmath52which indeed lies on a linear pomeron - type trajectory . the condition @xmath53 fixes the normalization constants@xmath54 from the general expression ( [ dc ] ) one then finds the glueball decay constants in the soft - wall background as @xcite @xmath55after fixing the factor @xmath56 by eq . ( [ est ] ) one can as above obtain quantitative estimates for the @xmath24 by setting the ir scale @xmath57 either such as to reproduce the typical quenched mass value @xmath58 1.5 gev or by using the value @xmath59 gev of ref . [ both variants lead to similar soft - wall predictions @xmath60 gev for the ground state decay constant for08 . holographic correlation functions can be derived from the on - shell action of the gravity dual which plays the role of their generating functional . to construct it , one employs the bulk - to - boundary propagator @xmath61 @xcite , i.e. the solution of the field equation ( [ fe ] ) subject to the @xmath62 and @xmath63 boundary conditions , to write the solution of eq . ( [ fe ] ) with a boundary source @xmath64 as @xmath65inserting the solution ( [ sln ] ) into eq . ( [ sd ] ) yields the on - shell action , and taking two functional derivatives with respect to @xmath66 then generates the correlator@xmath67 _ { z=\varepsilon \rightarrow 0 } \label{corr}\]]of the scalar glueball . analytical solutions for @xmath68 in both hard- and soft - wall backgrounds were found in ref . [ 6 ] . after plugging the analytical hard - wall solution for @xmath61 @xcite ( subject to the neumann ir boundary condition ) into the general expression ( [ corr ] ) and discarding two contact terms , one ends up with the correlator @xcite@xmath69 \label{hwc}\]](@xmath70 are mcdonald functions @xcite ) at spacelike momenta @xmath71 . its spectral density @xmath72 is defined by means of the dispersion relation@xmath73(suppressing again subtraction terms ) and takes the form @xcite@xmath74where the hard - wall mass spectrum @xmath75 @xmath76 reappears . the spectral weight ( [ hwsd ] ) is non - negative and consists of zero - width poles , as expected in the large-@xmath77 limit where glueballs are stable against strong decay . the overall correlator normalization @xmath78 is fixed by matching the leading conformal logarithm to the free qcd gluon loop,@xmath79with @xmath80 . for @xmath81 the holographic correlator ( hwc ) can be compared to the qcd short - distance expansion @xcite . the exponential @xmath82 dependence ( times powers of @xmath82 ) of its non - conformal part@xmath83 q^{4}e^{-2qz_{m } } \label{nph}\]]has its qcd ope counterpart in the small - size instanton contribution for01 * * * * @xmath84to the unit operator coefficient . since the instanton - induced correlations are attractive and of relatively short range @xmath85 , they reduce the scalar glueball mass and size while increasing its decay constant for01 . approximately equating eq . ( [ npi ] ) to the second term in eq . ( [ nph ] ) yields the holographic estimates * * * * @xmath86for the average instanton size @xmath87 and the overall instanton density @xmath88 in terms of the ir scale @xmath26 . the first relation reflects the duality between gauge instantons of size @xmath89 and pointlike @xmath90 instantons localized at @xmath91 . with @xmath92 gev it implies @xmath93 fm , i.e. almost twice the standard value @xcite . this may suggest that the bulk dynamics ( [ sd ] ) is more suitable for pure yang - mills theory with it larger instanton sizes @xmath94 fm , and that the strongly coupled hard - wall uv dynamics describes the small - instanton physics beyond the conformal regime rather poorly . the above discussion implies that @xmath95 , i.e. that large instantons , which would contribute to the condensates , are absent since their duals do not fit into the ads@xmath96 slice . hence the hard wall reproduces the qcd result that the _ hard _ nonperturbative physics from small instantons ( instead of the soft condensate physics ) dominates the short - distance @xmath23 glueball correlator @xcite . moreover , the qcd low - energy theorem @xcite @xmath97with @xmath98 is trivially satisfied by the hard - wall correlator ( [ hwc ] ) which vanishes at @xmath99 . this is consistent with eq . ( [ let ] ) since the gluon condensate vanishes in the hard wall as well . inserting the analytical solution for the soft - wall bulk - to - boundary propagator @xmath100 @xcite into eq . ( [ corr ] ) yields ( after discarding two divergent contact terms ) the soft - wall correlator @xcite@xmath101 \label{swc}\]]in terms of the digamma function @xmath102 @xcite . the analyticity structure of eq . ( [ swc ] ) implies that its spectral density has the form @xcite@xmath103which is non - negative for @xmath104 and consists of zero - width poles , as expected at large @xmath77 , at the soft - wall masses ( [ msw ] ) with residua determined by the soft - wall decay constants ( [ fsw ] ) . in order to compare the soft - wall correlator ( [ swc ] ) to the qcd ope , we use the asymptotic expansion of the digamma function to rewrite eq . ( swc ) for @xmath105 as @xmath106 \label{c}\]]which is renormalized at the ope scale @xmath107 . the normalization @xmath108 is again fixed by eq . ( [ est ] ) since large momenta @xmath109 probe the @xmath110 region where hard- and soft - wall correlators are governed by the same ads@xmath96-induced logarithm . besides the leading conformal and a second logarithmic term , the expansion ( [ c ] ) contains an infinite tower of power corrections . comparison with the ope suggests those to be related to the gauge - theory condensates @xmath111 of @xmath112 dimensional composite operators . for a first order - of - magnitude check one may equate the coefficients of the @xmath113 and @xmath114 terms in eq . ( [ c ] ) to the ( @xmath115 ) qcd wilson coefficients , yielding@xmath116for @xmath117 these are the rough magnitudes of the qcd condensates , but the sign of the qcd gluon condensate @xmath118 gev@xmath119 is positive . while qcd estimates of both signs exist for the three - gluon condensate , the above signs would also be at odds with the factorization approximation @xmath120 . the probably most intriguing prediction of the soft - wall correlator ( [ c ] ) is the additional power correction of dimension two which can not appear in the ope since qcd lacks a corresponding local operator . when linear contributions to the _ short_-distance heavy - quark potential are approximately described by a tachyonic gluon mass @xmath121 , however , one finds the correction @xcite@xmath122which has precisely the form of the second term in eq . ( [ c ] ) . comparison of the coefficients provides the holographic estimate @xmath123 and with @xmath124 further @xmath125 gev@xmath126 which is of the expected magnitude @xcite but again of opposite sign , i.e. not tachyonic . hence the expansion ( [ c ] ) completely reproduces the qualitative @xmath82 dependence of the qcd short - distance correlator ( to leading order in @xmath127 ) but fails to predict the sign of at least the two leading terms . this pattern can be better understood by recalling that the dimensions of the qcd condensates are generated by operators which are renormalized at @xmath128 gev and thus ir dominated . the form and general @xmath82 dependence of the qcd power corrections is therefore governed by ir physics , which may explain why the strongly - coupled soft - wall dynamics can reproduce it . the deviations of size and signs of the holographic power corrections from their qcd counterparts ( and the absence of radiative corrections ) should then originate mainly from the poorer description of the weakly - coupled , perturbative wilson coefficients by the soft - wall dynamics which lacks @xmath129 corrections and remains strongly coupled in the uv . the above interpretation also suggests an approximate separation of the soft - wall power corrections into wilson coefficients and condensates . indeed , under the premise that the strongly - coupled soft wall dynamics approximately reproduces the qcd condensate values , one may obtain holographic estimates for the wilson coefficients the gluon condensate coefficient , e.g. , becomes with @xmath130 @xcite and @xmath131 @xcite @xmath132its smaller size and opposite sign relative to the qcd result provides some intuition for the soft - wall deficiencies in describing weakly - coupled qcd physics . of course , the estimate ( [ cg2 ] ) is prone to further error sources , including uncertainties in the qcd condensate values and their sensitivity to light quark contributions . ( the above separation into hard and soft contributions would fail for the two - dimensional power correction , incidentally , since both the gluon mass @xmath121 and its coefficient receive uv contributions . ) since the condensates are hadron - channel independent while the wilson coefficients are not , one would further expect to obtain inconsistent condensate estimates when trying to extract them in different channels by relying on the respective qcd wilson coefficients . finally , we note that the soft - wall correlator ( [ swc ] ) with @xmath133 violates the low - energy theorem ( [ let ] ) since eq . ( [ g4 ] ) implies a finite rhs . recently an attempt has been made to generalize the soft - wall glueball correlator ( [ swc ] ) by adding to the bulk - to - boundary propagator @xmath134 the at small @xmath20 subleading solution of eq . ( [ fe ] ) , multiplied by an a priori arbitrary coefficient function @xmath135 @xcite . the added solution blows up at large @xmath20 , i.e. it violates the standard `` regularity in the bulk '' condition @xcite and requires an ad - hoc ir cutoff prescription without obvious correspondence on the gauge - theory side ( in contrast to the standard uv renormalization whose `` dual '' tames the volume divergence of the on - shell bulk action ) . the resulting expression for the correlator differs from ours , i.e. eq . ( swc ) , by the addition of the arbitrary coefficient function @xmath136 . any desired behavior of the correlator could thus be chosen by hand , and independently of the soft - wall background , by adapting @xmath137 accordingly . [ 12 ] attempted to use this apparent freedom to equate the @xmath138 power corrections to their qcdvalues and to eliminate the two - dimensional power correction . ) hence the prescription of ref . [ 12 ] gives up on the one - to - one duality between the gauge vacuum and the gravity background ( together with other parts of the ads / cft dictionary ) and results in a practically total loss of predictive power . moreover , it is internally inconsistent . indeed , it was shown to result in the same mass spectrum ( [ msw ] ) and the same decay constants ( [ fsw ] ) as in our case @xcite . hence it must reproduce our spectral density ( [ swsd ] ) and thus the physics content of the correlator ( [ swc ] ) , i.e. any remaining discrepancies have to originate from the unphysical contact terms which are needed to regularize the dispersion integral ( [ drel ] ) . this is in direct contradiction to the supposed freedom of adding an arbitrary function @xmath135 to the correlator . the loss of predictivity , uniqueness and consistency incurred when relaxing the regularity of @xmath68 in the bulk has a common origin . indeed , any given ( smooth ) function defined on the uv boundary @xmath139 of ( euclidean and compactified ) ads@xmath140 is known to have a _ unique _ extension to a solution of the massless scalar field equation in the ads@xmath140 bulk @xcite . when applied to the boundary source @xmath141 , this mathematical fact ensures the one - to - one correspondence between the gauge - theory operators and the dual string mode solutions . the inconsistency of the attempted generalization of the bulk - to - boundary propagator arises from the violation of this fact . in order to maintain the mathematically required uniqueness of the relation ( [ sln ] ) between a given boundary source @xmath142 and its dual mode solution @xmath143 , the uv - subleading solution must not be added to @xmath68 . we have derived and analyzed the predictions of the two currently most popular ads / qcd duals , i.e. the hard - wall and dilaton soft - wall backgrounds , for the @xmath23 glueball correlation function and decay constants . in their representation of specific nonperturbative glueball physics ( at momenta larger than the qcd scale ) both holographic duals turn out to complement each other : the soft - wall correlator contains all known types of qcd power corrections , generated both by vacuum condensates and by a hypothetical uv gluon mass , while sizeable exponential corrections of the type induced by small - scale qcd instantons are reproduced in the hard - wall correlator . since the qcd power corrections to the @xmath23 glueball correlator are suppressed by unusually small wilson coefficients whereas the small - instanton contributions are enhanced , the hard - wall background may provide a more reliable approximation to the scalar glueball correlator . furthermore , the above complementarity , which helps to relate holographic predictions ( and their limitations ) to specific aspects of the gauge dynamics , should extend to other hadron channels . while all our holographic estimates have the order of magnitude expected from qcd@xmath144 the signs of the two leading soft - wall power corrections are opposite to those of standard qcd estimates ( and in conflict with the factorization approximation for the four - gluon condensate ) . we have argued that this provides evidence for the short - distance physics in the wilson coefficients to be inadequately reproduced by the strongly - coupled uv dynamics of bottom - up models ( beyond the leading conformal logarithm ) . we have further shown that this problem can not be mended by admixing the uv - subleading solution to the bulk - to - boundary propagator ( as recently advocated ) without loosing consistency and predictive power . in addition , we have provided first holographic estimates for the @xmath23 glueball decay constants which contain glueball size information , are important for experimental glueball searches and probe aspects of the dual dynamics to which the mass spectrum is less sensitive . the hard - wall prediction for the ground - state decay constant @xmath145 is more than twice as large as its soft - wall counterpart . this is a consequence of the exponential contributions to the hard - wall correlator which reproduce the strong instanton - induced short - distance attraction in the qcd correlator . the hard - wall prediction @xmath146 gev implies an exceptionally small glueball size and agrees inside errors with iope sum - rule and lattice results . we would like to thank pietro colangelo and his coworkers for correspondence and acknowledge financial support from the fundao de amparo a pesquisa do estado de so paulo ( fapesp ) and the deutsche forschungsgemeinschaft ( dfg ) . k. peeters and m. zamaklar , eur . j. special topics * 152 * , 113 ( 2007 ) ; s.j . brodsky and g.f . de tramond , arxiv:0802.0514 ; j. erdmenger , n. evans , i. kirsch and e. threlfall , eur . phys . j. a * 35 * , 81 ( 2008 ) . grigoryan and a.v . radyushkin , phys . b * 650 * , 421 ( 2007 ) ; phys . d * 76 * , 095007 ( 2007 ) ; phys . d * 76 * , 115007 ( 2007 ) ; phys . d * 77 * , 115024 ( 2008 ) ; s.j . brodsky and g.f . de tramond , phys . d. * 77 * , 056007 ( 2008 ) ; arxiv:0804.0452 ; h.j . kwee and r.f . lebed , jhep * 0801 * , 027 ( 2008 ) ; d.k . hong , m. rho , h .- u . yi , phys . d * 77 * , 014030 ( 2008 ) ; k. hashimoto , t. sakai and s. sugimoto , arxiv:0806.3122 ; k.y . kim and i. zahed , arxiv:0807.0033 . y. chen et al . d * 73 * , 014516 ( 2006 ) ; m. loan and y. ying , hep - lat/0603030 ; n. ishii , h. suganuma and h. matsufuru , phys . rev . d * 66 * , 94506 ( 2002 ) ; p. de forcrand and k liu , phys . * 69 * , 245 ( 1992 ) ; r. gupta et al . d * 43 * , 2301 ( 1991 ) .	we investigate the dynamical content of both hard- and soft - wall approximations to holographic qcd by deriving the corresponding glueball correlation functions and by confronting them with a variety of qcd results . we further calculate the glueball decay constants in both holographic duals , discuss emerging limitations and improvement strategies , and comment on a recent attempt to generalize the glueball correlator in the soft - wall background .
new physics that affects the observability of the higgs boson of the standard model ( sm ) is of utmost importance to study . one the simplest kinds of new physics is a sequential replication of the three generations of chiral matter @xcite . such a fourth generation has been considered and forgotten or discarded many times , wrongly leaving the impression that it is either ruled out or highly disfavored by experimental data ( for instance , see ref . @xcite ) . the status of four generations is more subtle @xcite . ref . @xcite analyzed the contributions of one ( and more ) extra generations to the oblique parameters and explicitly found that one generation can be perfectly consistent with a heavy ( @xmath2 gev ) higgs . these significant results are primarily based on numerical scans , with emphasis on the role of a lighter neutrino ( @xmath3 gev ) to minimize the contributions to the oblique parameters ( see also ref . however , a neutrino with mass of @xmath3 gev , if unstable , is ruled out by lep ii bounds , while if it exactly stable , may be ruled out by dark matter direct search experiments @xcite . correlations of the mass parameters leading to viable spectra are certainly not transparent , making it hard to determine how to parse their results against present experimental bounds . subsequent analyses @xcite studied the relationships among fourth generation parameters , but their analysis was performed using a global ( numerical ) fit to 2001 electroweak data and again emphasized a 50 gev neutrino . electroweak data has since been refined ( in particularly @xmath4 ) , so these results no longer obviously apply , in particular if we incorporate a heavier neutrino . the impact of a chiral fourth generation on higgs physics has been briefly discussed @xcite , however the range of masses that were considered were not necessarily correlated to the fourth generation mass spectra and higgs mass appropriate to satisfy current direct search bounds and electroweak precision constraints . moreover , in cases where there is overlap , our results do not always agree ; we point out the differences below . in this paper , we first systematically determine the allowed parameter space of fourth generation masses and mixings . we find quite simple mass relations that minimize the precision electroweak oblique parameters , so our analysis can easily be extended to future refinements in electroweak measurements . we then use typical spectra to compute the consequences for fourth generation particle production and decay , as well as the effects on the higgs sector of the standard model . we find that a wide range of higgs masses is consistent with electroweak data , leading to significant modifications of higgs production and decay . we outline the major effects , identifying the well - known effects from others that ( to our knowledge ) are new . there are in addition spectacular signals of the fourth generation itself . given that direct searches at lep ii and tevatron have already constrained the masses somewhat , we can expect future searches at tevatron will continue to push the limits up , but will not rule out four generations . the lhc is able probe heavy quarks throughout their mass range . many of the signals have been recently considered ( albeit in somewhat different mass ranges and context from what we consider here ) in refs . @xcite , to which we refer the interested reader . the framework we consider is to enlarge the standard model to include a complete sequential fourth generation of chiral matter ( @xmath5 , @xmath6 , @xmath7 , @xmath8 , @xmath9 ) as well as a single right - handed neutrino @xmath10 . yukawa couplings and right - handed neutrino masses are given by @xmath11 the generation indices are @xmath12 while we reserve @xmath13 for the standard model . su(2 ) contractions are implicit . light neutrino masses can arise from either a hierarchy in neutrino yukawa couplings @xmath14 or right - handed neutrino masses @xmath15 or some combination . ( for an amusing combination , see ref . we mainly consider two possibilities for the fourth generation neutrino mass : purely dirac ( @xmath16 ) and mixed ( @xmath17 ) . there are four obvious restrictions on a fourth generation : ( 1 ) the invisible width of the @xmath18 ; ( 2 ) direct search bounds ; ( 3 ) generational mixing ; ( 4 ) oblique electroweak effects . we now discuss them one - by - one . once a fourth generation neutrino has a mass @xmath19 , the constraint from the invisible @xmath18 width becomes moot . assuming non - zero mixings @xmath20 or @xmath21 , the fourth generation quarks , charged lepton , and neutrino decay , and thus there are no cosmological constraints from stable matter . ( we will briefly comment on neutrino dark matter at the end the paper . ) a robust lower bound on fourth generation masses comes from lep ii . the bound on unstable charged leptons is @xmath22 gev , while the bound on unstable neutral dirac neutrinos is @xmath23 gev for the decay modes @xmath24 . these limits are weakened only by about 10 gev when the neutrino has a majorana mass . because the small differences in the bounds between different flavors , charged versus neutral leptons , and majorana versus dirac mass do not affect our results , we apply the lep ii bounds as @xmath25 gev throughout . the tevatron has significantly greater sensitivity for fourth generation quarks @xcite . the strongest bound is from the cdf search for @xmath26 , obtaining the lower bound @xmath27 gev to 95% confidence level ( cl ) . no @xmath28-tag was used , so there is no dependence on the final state jet flavor , and hence this limit applies independent of the ckm elements @xmath29 . there is no analogous limit on the mass of @xmath7 . if @xmath30 and @xmath31 , then the dramatic @xmath32 signal may be confused into the top sample . if the decay proceeds through a lighter generation , then the production rate and signal are the same as for @xmath6 , and so we expect a bound on the mass of @xmath7 similar to that on @xmath6 . if @xmath33 , then @xmath7 decay could proceed through a `` doubly - cabbibo '' suppressed tree level process @xmath34 or through the one - loop process @xmath35 . the relative branching ratios depend on details @xcite . in particular , taking @xmath36 , cdf obtains the bound @xmath37 gev at 95% cl @xcite we choose to adopt the largely ckm - independent bound @xmath38 gev throughout . the off - diagonal elements @xmath39 of the ckm matrix @xmath40 are constrained by flavor physics . as in the standard model , the flavor - violating neutral current effects occur in loops and are automatically gim suppressed . rough constraints on the mixing between the first / second and fourth generation can be extracted requiring unitarity of the enlarged @xmath41 ckm matrix . the sm @xmath42 sub - matrix is well tested by a variety of sm processes @xcite . the first row of the matrix , combined with measurements of @xmath43 , @xmath44 , and @xmath45 , yields @xmath46 for the second row we can use the hadronic @xmath47 branching ratio to obtain @xmath48 similarly , the first column of the matrix allows one to infer , @xmath49 if we require the above relations be satisfied to @xmath50 , we obtain @xmath51,@xmath52 . the remainder of the elements ( @xmath53 , @xmath54 , @xmath55 , and @xmath56 ) could be constrained through a global fit to the @xmath57 ckm matrix , including the contributions of the fourth generation quarks to specific observables in loops ( for example @xcite ) , but this is beyond the scope of this work . similarly there are two additional cp - violating phases in the @xmath41 ckm matrix , but since their effects are proportional to the unknown ( real parts ) of the off - diagonal ckm mixings , we ignore their effects . the least constrained sector is the mixing between the third and fourth generations . the observation of single top production @xcite can be used to obtain a lower limit @xmath58 at @xmath1 c.l . @xcite , which still allows for large third / fourth generation mixing . thus it seems likely that fourth generation charged - current decays will be mostly into third generation quarks , provided the mass difference is large enough to permit two - body decays . the new elements in the pmns matrix @xmath59 also have constraints from lepton flavor violation in the charged and neutral sectors . the most stringent constraint is the absence of @xmath60 . for weak scale purely dirac neutrinos this constraint @xcite implies @xmath61 . this suggests that first / second generation mixings with the fourth generation should be smaller than about @xmath62 . other generational mixings can also be constrained from the absence of lepton flavor violating effects , where again third / fourth generation mixings are ( as expected ) the most weakly constrained . there is , however , a significant constraint from neutrinoless double beta decay on @xmath63 in the presence of a weak scale majorana mass @xmath64 . such a decay can be mediated by a very light neutrino mixing with a weak scale ( partly ) majorana neutrino . using ref . @xcite and assuming only first / fourth generational mixing , we obtain @xmath65 where @xmath66 and pmns phases are ignored . this expression is valid as long as the fourth generation neutrino masses exceed the characteristic energy scale of the double beta nuclear process , @xmath67 mev . inserting characteristic values , we obtain @xmath68 no bound remains once the fourth generation majorana mass is made small , @xmath69 mev . the most pernicious effect of a fourth generation is the contribution to oblique electroweak corrections . @xmath70 mixing is enhanced , leading to a positive contribution @xmath71 in the limit of degenerate isospin multiplets ( quark and lepton ) . degeneracy is usually assumed for simplicity since split doublets significantly contribute to the isospin violating parameter @xmath72 . there are three important effects that can mitigate the contribution to @xmath73 . the first , and most important , is exploiting the relative experimental insensitivity to the @xmath74 direction in oblique parameter space . we will be more precise below , but suffice to say slightly split electroweak doublets are in far better agreement with electroweak data than without the @xmath75 contribution . the second effect involves a reduced contribution to @xmath76 by splitting the fourth generation multiplets in a particular mass hierarchy . the last , and least important effect is introducing a majorana mass for the fourth generation neutrino . splitting the up - type from down - type fermion masses in the same electroweak doublet can give a negative contribution to @xmath76 . in the large mass limit @xmath77 , the contribution to @xmath76 depends logarithmically on the ratio @xmath78 @xcite @xmath79 where @xmath80 is the hypercharge of the left - handed doublet of fermions with degeneracy ( color factor ) @xmath81 . clearly the fourth generation contributions to @xmath76 are reduced if @xmath82 for quarks ( @xmath83 ) and @xmath84 for leptons @xmath85 . how big can this effect be given that split multiplets also contribute to @xmath75 ? ( diagonal , blue ) and @xmath86 ( horizontal , red ) for the fourth generation quarks . the plot is symmetric with respect to @xmath87 , since @xmath86 is positive definite . the tevatron bound @xmath38 gev excludes the shaded ( yellow ) region . ] to calculate @xmath73 ( and @xmath75 and @xmath88 ) we use exact one - loop expressions which are valid for all @xmath89 @xcite . we checked our formulae by explicitly verifying finiteness ( renormalization scale independence ) as well as finding numerical agreement with several explicit results given in ref . @xcite . in fig . [ quarksst - fig ] we show the size of the contribution from the @xmath90 doublet as a function of the masses of the quarks . the effect of using the exact one loop expressions is modest ; in fact eq . ( [ s - eq ] ) reproduces the @xmath76 contours up to an accuracy @xmath91 throughout the plot . the typical size of @xmath92 is smaller than @xmath62 everywhere , and so we set @xmath93 throughout . for the leptons , what is most important is the split between the neutral and charged fermion masses . for example , @xmath94 gev implies @xmath95 , and the slightly larger values @xmath96 gev give @xmath97 . these results from the exact one - loop formulae agree surprisingly well with eq . ( [ s - eq ] ) , despite the lepton masses being near @xmath98 . fits of the combined electroweak data provide constraints on the oblique parameters and have been performed by the lep electroweak working group ( lep ewwg ) @xcite and separately by the pdg @xcite . both fits find that the standard model defined by @xmath99 with @xmath100 gev and @xmath101 gev is within @xmath50 of the central value ( always holding @xmath93 ) . however , the two fits disagree on the best fit point . the latest lep ewwg fit finds a central value @xmath102 @xcite with a @xmath0 contour that is elongated along the @xmath103 major axis from @xmath104 to @xmath105 . by contrast , the pdg find the central value @xmath106 after adjusting @xmath72 up by @xmath107 to account for the latest value of @xmath100 gev . the most precise constraints on @xmath76 and @xmath72 arise from @xmath108 and @xmath4 , used by both groups . the actual numerical constraints derived from these measurements differ slightly between each group , presumably due to slight updates of data ( the @xmath76-@xmath72 plot generated by the 2006 lep ewwg is one year newer than the plot included in the 2006 pdg ) . a larger difference concerns the use of the @xmath18 partial widths and @xmath109 . the lep ewwg advocate using just @xmath110 , since it is insensitive to @xmath111 . this leads to a flatter constraint in the @xmath76-@xmath72 plane . the pdg include the @xmath111-sensitive quantities @xmath112 , @xmath109 , @xmath113 as well as @xmath114 , and obtain a less flat , more oval - shaped constraint . additional lower energy data can also be used to ( much more weakly ) constrain @xmath76 and @xmath72 , although there are systematic uncertainties ( and some persistent discrepancies in the measurements themselves ) . the lep ewwg do not include lower energy data in their fit , whereas the pdg appear to include some of it . in light of these subtleties , we choose to use the lep ewwg results when quoting levels of confidence of our calculated shifts in the @xmath76-@xmath72 plane . we remind the reader , however , that the actual level of confidence is obviously a sensitive function of the precise nature of the fit to electroweak data . .examples of the total contributions to @xmath73 and @xmath75 from a fourth generation . the lepton masses are fixed to @xmath115 gev and @xmath116 gev , giving @xmath117 and @xmath118 . the best fit to data is @xmath102 @xcite . the standard model is normalized to @xmath119 for @xmath100 gev and @xmath101 gev . all points are within the 68% cl contour defined by the lep ewwg @xcite . [ cols="^,^,^,^,^,^",options="header " , ] finally , higgs pair production is resurrected by fourth generation loop effects . while the sm production rate at the lhc might barely be sufficient to confirm the existence of a triple higgs coupling @xmath120 as predicted by the higgs potential @xcite , the enhancement of the effective @xmath121 and @xmath122 couplings should allow for a proper measurement of @xmath120 . enhancements to higgs pair production using an operator approach was also recently considered in ref . @xcite . total rates are notoriously difficult observables at hadron colliders , but the higgs self coupling can be beautifully extracted from the threshold behavior of the @xmath123 amplitude . at threshold , this process is dominated by the two form factors @xmath124 proportional to the metric tensor , which arise from the triangular and box diagrams ( following the notation of ref . @xcite ) . in the low energy limit @xcite the box diagram s form factor proportional to the transverse tensor is suppressed by powers of the loop mass . the higgs coupling analysis makes use of the fact that at threshold the two contributions @xmath125 and @xmath126 cancel . more precisely , in the low energy limit @xmath127 we find @xmath128 . this cancellation explains the increase in rate when we set @xmath120 to zero , as shown in table [ tab : pairs ] . if we only slightly vary the size of the higgs self coupling , this threshold behavior changes significantly @xcite and provides an experimental handle on @xmath120 . in figure [ dihiggs - fig ] we show the @xmath129 invariant mass ( or @xmath130 at parton level ) distribution . the shift between finite and zero @xmath120 in the standard model provides the ( s)lhc measurement of the higgs self coupling . similarly to the @xmath121 form factors shown in table [ tab : formfactors ] the decoupling assumption for top quarks is numerically not quite as good as for the additional fourth generation quarks . once the process is dominated by heavier quarks the variation of @xmath131 with @xmath120 becomes significantly more pronounced , so there is little doubt that we can use it to measure the higgs self coupling . for the standard model , the higgs self coupling analysis at the lhc is likely restricted to the @xmath132 decay channel @xcite . from table [ tab : pairs ] we see that for light higgs masses this decay is strongly suppressed , so it would be an interesting exercise to see if there are alternative decay channels @xcite which might work for lighter higgs bosons , given the rate and @xmath131 sensitivity increase by the fourth generation . until now we have concentrated on collider effects of a fourth generation coupled to one higgs doublet . since the yukawa couplings of the new fermions exceed @xmath133 for the fourth generation quarks , the four generation model as an effective theory breaks down at a scale that may not be far above the tev scale . there are two well - known constraints : ( 1 ) the possibility that the quartic coupling is driven negative , destabilizing the electroweak scale by producing large field minima through quantum corrections @xcite , and ( 2 ) large yukawa couplings driving the higgs quartic and/or the yukawas themselves to a landau pole , _ i.e. _ entering a strong coupling regime . in both cases the problematic coupling is the higgs quartic , since it receives much larger new contributions to its renormalization group running from the fourth generation quark yukawas couplings . the renormalization group equation for @xmath134 is @xmath135 where we have shown only the dominant terms . the last two terms encode the higgs wave function and quartic terms induced by the fermions ; the sum is over all identical fermions with degeneracy @xmath136 . in our numerical estimations we also include the sub - leading electroweak coupling dependence , and evolve using the full set of one loop @xmath137-functions @xcite . we can estimate the scale at which the meta - stability bound becomes problematic by requiring that the probability of tunneling into another vacuum over the current age of the universe is much less than 1 . this is equivalent to the requirement that the running quartic interaction is @xcite @xmath138 where @xmath139 is the hubble scale . the scale at which this inequality is saturated is a minimum scale where new physics is required . we should emphasize that the new physics does _ not _ need to be strongly coupled . for example , a supersymmetric model with four generations does not have a running quartic that turns negative as long as superpartners are ( roughly ) below the tev scale . this is important because weakly coupled physics with particles obtaining their mass through _ e.g. _ supersymmetry breaking , not electroweak breaking , will hardly affect our higgs results . . these two constraints are qualitatively distinct : meta stability can be restored by _ weakly _ coupled physics below a tev scale , whereas the landau pole signals a strongly interacting higgs sector . the dashed curve reproduces the sm triviality bound . ] the second constraint is potentially a stronger one . requiring that the quartic remain perturbative , @xmath140 , we find that the upper bound on the cutoff scale of the theory rapidly becomes small as the higgs mass is increased . we show this constraint as well as the meta - stability constraint in fig . [ stability - fig ] . we find that for our choices of fourth generation masses , the yukawa interactions remain perturbative to slightly beyond the higgs meta - stability / triviality bounds for all considered higgs masses . the `` chimney '' region , in which the effective theory of the standard model with @xmath141 gev remains valid to @xmath142 , closes off . we find the maximal cutoff scale before new physics of any kind enters occurs for higgs masses in the neighborhood of 300 gev . much lower higgs masses , in particular @xmath143 , imply other new physics must enter to prevent developing a deeper minimum away from the electroweak breaking vacuum . nevertheless , we emphasize that this new physics can be weakly coupled below a tev with little effect on higgs physics itself . conversely , to resolve the physics of the cutoff scale in the case where the quartic ( or the yukawas ) encounter a landau pole undoubtedly requires physics directly connected to electroweak symmetry breaking . this new physics could be stronger - coupled supersymmetry , technicolor , topcolor , or a little higgs construction . we have considered the constraints on a fourth generation and its effects on higgs physics in the standard model . if nature does indeed have a fourth generation , it is amusing to speculate on the rich series of new phenomena expected at colliders now operating and about to begin . the ordering of discoveries could proceed by tevatron discovering the higgs , with an unusually large production cross section , or in mass range that was previously thought to be undetectable in the standard model . subdominant decays of the higgs may reveal a new sector . direct production of fourth generation neutrinos or leptons may also be possible at tevatron , but relies on a more detailed understanding the background . once the lhc turns on , the fourth generation quarks should be readily produced and found . the higgs can be found using the golden mode for a wide range of mass , and for most of this range , it will be found very quickly with a small integrated luminosity ( due to the large enhancement of the gluon fusion channel ) . given measures of the cross section for higgs production as well as branching ratios of higgs into subdominant modes , the lhc will be able to rapidly verify that a fourth _ chiral _ generation does indeed exist . while our focus has been on the _ effects _ of a fourth generation , there is also the possibility that a fourth generation could alleviate or solve some of the pressing problems addressed by other models of new physics . one amusing possibility is to employ a variation of the mechanism of ref . @xcite to revive electroweak baryogenesis in the ( four - generation ) standard model . another possibility is to impose a parity symmetry to stabilize the fourth generation lepton to serve as cold dark matter . this is naively ruled out by direct detection , however there are mechanisms @xcite to avoid these bounds by either splitting the neutrino eigenstates with a small majorana mass or otherwise invoking additional physics such as a @xmath144 coupling to @xmath145 . a detailed study of these issues is in progress and will be reported on elsewhere . we thank d.e . kaplan , e. katz , m. schmaltz , and j. wacker for discussions and c. wagner for reminding us of the true and electroweak baryogenesis . gdk , tp , and tt thank the aspen center of physics for a stimulating atmosphere where this work was begun . gdk thanks supa and argonne national laboratory for hospitality where part of this work was completed . this work was supported in part by the department of energy under contracts de - fg02 - 96er40969 ( gdk ) and de - ac02 - 06ch11357 ( tt ) . one could also speculate about five or more generations . optimizing over particle masses , we find that the contribution to the oblique electroweak parameters is @xmath146 and thus outside the 95% cl ellipse . w. m. yao _ et al . _ [ particle data group ] , j. phys . g * 33 * , 1 ( 2006 ) . p. h. frampton , p. q. hung and m. sher , phys . * 330 * , 263 ( 2000 ) [ arxiv : hep - ph/9903387 ] . h. j. he , n. polonsky and s. f. su , phys . d * 64 * , 053004 ( 2001 ) [ arxiv : hep - ph/0102144 ] . m. maltoni , v. a. novikov , l. b. okun , a. n. rozanov and m. i. vysotsky , phys . b * 476 * , 107 ( 2000 ) [ arxiv : hep - ph/9911535 ] . direct search experiments have ruled out a stable fourth generation dirac neutrino with a cosmological relic density ( as well as densities much lower than even this ) . whether the thermal relic abundance can be arranged to sufficiently small so as to contribute insignificantly to cosmological dark matter ( and thus avoid the direct search bounds ) depends on how large the annihilation cross section can be made , which in turn depends on several parameters including the majorana mass of the neutrino . see for a recent discussion . v. a. novikov , l. b. okun , a. n. rozanov and m. i. vysotsky , phys . b * 529 * , 111 ( 2002 ) [ arxiv : hep - ph/0111028 ] . v. a. novikov , l. b. okun , a. n. rozanov and m. i. vysotsky , jetp lett . * 76 * , 127 ( 2002 ) [ pisma zh . . fiz . * 76 * , 158 ( 2002 ) ] [ arxiv : hep - ph/0203132 ] . e. arik , m. arik , s. a. cetin , t. conka , a. mailov and s. sultansoy , eur . j. c * 26 * , 9 ( 2002 ) [ arxiv : hep - ph/0109037 ] . e. arik , o. cakir , s. a. cetin and s. sultansoy , phys . d * 66 * , 033003 ( 2002 ) [ arxiv : hep - ph/0203257 ] . e. arik , o. cakir , s. a. cetin and s. sultansoy , acta phys . b * 37 * , 2839 ( 2006 ) [ arxiv : hep - ph/0502050 ] . b. holdom , jhep * 0608 * , 076 ( 2006 ) [ arxiv : hep - ph/0606146 ] ; b. holdom , jhep * 0703 * , 063 ( 2007 ) [ arxiv : hep - ph/0702037 ] ; b. holdom , arxiv:0705.1736 [ hep - ph ] . w. skiba and d. tucker - smith , phys . rev . d * 75 * , 115010 ( 2007 ) [ arxiv : hep - ph/0701247 ] . c. t. hill and e. a. paschos , phys . b * 241 * , 96 ( 1990 ) . the tevatron also has sensitivity to the charged and neutral leptons , although there are several subtleties involving triggering ( particularly when @xmath147 is small ) and backgrounds . we are not aware of robust bounds from run ii data ( independently of pmns flavor mixing , for example ) . however , it may be possible to eliminate some lepton mass parameter space for some or all flavor - mixings with a careful analysis . ` http://www-cdf.fnal.gov/physics/new/top/2005/ ` ` ljets / tprime / gen6/public.html ` m. sher , phys . d * 61 * , 057303 ( 2000 ) [ arxiv : hep - ph/9908238 ] . a. arhrib and w. s. hou , phys . d * 64 * , 073016 ( 2001 ) [ arxiv : hep - ph/0012027 ] . cdf collaboration , arxiv:0706.3264 [ hep - ph ] . w. s. hou , r. s. willey and a. soni , phys . lett . * 58 * , 1608 ( 1987 ) [ erratum - ibid . * 60 * , 2337 ( 1988 ) ] ; w. s. hou , a. soni and h. steger , phys . b * 192 * , 441 ( 1987 ) . v. m. abazov _ et al . _ [ d0 collaboration ] , arxiv : hep - ex/0612052 . cdf collaboration , preliminary conference note 8588 ; ` http://www-cdf.fnal.gov/ ` ` physics / new / top/2006/singletop / me_1fb / index.html ` t. p. cheng and l. f. li , _ oxford , uk : clarendon ( 1984 ) 536 p. ( oxford science publications ) _ p. bamert , c. p. burgess and r. n. mohapatra , nucl . phys . b * 438 * , 3 ( 1995 ) [ arxiv : hep - ph/9408367 ] . m. e. peskin and t. takeuchi , phys . d * 46 * , 381 ( 1992 ) . b. a. kniehl , nucl . b * 352 * , 1 ( 1991 ) . j. alcaraz _ et al . _ [ aleph collaboration ] , arxiv : hep - ex/0612034 . the latest lep ewwg fit to @xmath148 can be found here : ` http://lepewwg.web.cern.ch/lepewwg/plots/ ` + ` summer2006/s06_stu_contours.eps ` the best fit to data is @xmath149 where the sm , defined by @xmath150 gev and @xmath151 gev , is at the origin @xmath99 . readjusting the definition of @xmath99 to @xmath100 gev @xcite and @xmath101 gev shifts the ellipse by an amount @xmath152 . e. gates and j. terning , phys . lett . * 67 * , 1840 ( 1991 ) . b. a. kniehl and h. g. kohrs , phys . d * 48 * , 225 ( 1993 ) . j. f. gunion , d. w. mckay and h. pois , phys . b * 334 * , 339 ( 1994 ) [ arxiv : hep - ph/9406249 ] ; j. f. gunion , d. w. mckay and h. pois , phys . d * 53 * , 1616 ( 1996 ) [ arxiv : hep - ph/9507323 ] . j. r. ellis , m. k. gaillard and d. v. nanopoulos , nucl . b * 106 * , 292 ( 1976 ) ; m. a. shifman , a. i. vainshtein , m. b. voloshin and v. i. zakharov , sov . j. nucl . phys . * 30 * , 711 ( 1979 ) [ yad . fiz . * 30 * , 1368 ( 1979 ) ] ; b. a. kniehl and m. spira , z. phys . c * 69 * , 77 ( 1995 ) [ arxiv : hep - ph/9505225 ] ; for an overview , see _ e.g. _ , a. djouadi , arxiv : hep - ph/0503172 . a. v. manohar and m. b. wise , phys . b * 636 * , 107 ( 2006 ) [ arxiv : hep - ph/0601212 ] . s. c. hsu [ cdf collaboration ] , arxiv:0706.2200 [ hep - ex ] . d0 collaboration , ` http://www-d0.fnal.gov/run2physics/ ` ` www / results / prelim / higgs / h21/h21f0_supplement.eps ` t. plehn , d. l. rainwater and d. zeppenfeld , phys . lett . b * 454 * , 297 ( 1999 ) [ arxiv : hep - ph/9902434 ] ; a. alves , o. eboli , t. plehn and d. l. rainwater , phys . d * 69 * , 075005 ( 2004 ) [ arxiv : hep - ph/0309042 ] . a. djouadi , j. kalinowski and m. spira , comput . commun . * 108 * , 56 ( 1998 ) [ arxiv : hep - ph/9704448 ] . our results differ from ref . @xcite who found a more modest reduction in the @xmath153 mode . d. l. rainwater and d. zeppenfeld , phys . d * 60 * , 113004 ( 1999 ) [ erratum - ibid . d * 61 * , 099901 ( 2000 ) ] [ arxiv : hep - ph/9906218 ] ; n. kauer , t. plehn , d. l. rainwater and d. zeppenfeld , phys . b * 503 * , 113 ( 2001 ) [ arxiv : hep - ph/0012351 ] . m. dittmar and h. k. dreiner , phys . d * 55 * , 167 ( 1997 ) [ arxiv : hep - ph/9608317 ] . v. bscher and k. jakobs , int . j. mod . a * 20 * , 2523 ( 2005 ) [ arxiv : hep - ph/0504099 ] . o. j. p. eboli and d. zeppenfeld , phys . b * 495 * , 147 ( 2000 ) [ arxiv : hep - ph/0009158 ] . t. plehn , d. l. rainwater and d. zeppenfeld , phys . rev . * 88 * , 051801 ( 2002 ) [ arxiv : hep - ph/0105325 ] ; v. hankele , g. klamke , d. zeppenfeld and t. figy , phys . d * 74 * , 095001 ( 2006 ) [ arxiv : hep - ph/0609075 ] . j. alwall _ et al . _ , arxiv:0706.2334 [ hep - ph ] . v. del duca , w. kilgore , c. oleari , c. r. schmidt and d. zeppenfeld , arxiv : hep - ph/0109147 . m. l. graesser , arxiv:0704.0438 [ hep - ph ] ; m. l. graesser , arxiv:0705.2190 [ hep - ph ] . u. baur , t. plehn and d. l. rainwater , phys . * 89 * , 151801 ( 2002 ) [ arxiv : hep - ph/0206024 ] ; u. baur , t. plehn and d. l. rainwater , phys . d * 67 * , 033003 ( 2003 ) [ arxiv : hep - ph/0211224 ] ; see also : f. gianotti _ et al . _ , eur . j. c * 39 * , 293 ( 2005 ) [ arxiv : hep - ph/0204087 ] . a. pierce , j. thaler and l. t. wang , arxiv : hep - ph/0609049 . t. plehn , m. spira and p. m. zerwas , nucl . b * 479 * , 46 ( 1996 ) [ erratum - ibid . b * 531 * , 655 ( 1998 ) ] [ arxiv : hep - ph/9603205 ] . u. baur , t. plehn and d. l. rainwater , phys . d * 69 * , 053004 ( 2004 ) [ arxiv : hep - ph/0310056 ] . m. sher , phys . rept . * 179 * , 273 ( 1989 ) . h. arason , d. j. castano , b. keszthelyi , s. mikaelian , e. j. piard , p. ramond and b. d. wright , phys . d * 46 * , 3945 ( 1992 ) . g. isidori , g. ridolfi and a. strumia , nucl . b * 609 * , 387 ( 2001 ) [ arxiv : hep - ph/0104016 ] . m. carena , a. megevand , m. quiros and c. e. m. wagner , nucl . b * 716 * , 319 ( 2005 ) [ arxiv : hep - ph/0410352 ] . d. r. smith and n. weiner , phys . d * 64 * , 043502 ( 2001 ) [ arxiv : hep - ph/0101138 ] . g. belanger , a. pukhov and g. servant , arxiv:0706.0526 [ hep - ph ] . g. servant and t. m. p. tait , new j. phys . * 4 * , 99 ( 2002 ) [ arxiv : hep - ph/0209262 ] . [ cdf collaboration ] , arxiv : hep - ex/0703034 .	in the light of the lhc , we revisit the implications of a fourth generation of chiral matter . we identify a specific ensemble of particle masses and mixings that are in agreement with all current experimental bounds as well as minimize the contributions to electroweak precision observables . higgs masses between 115 - 315 ( 115 - 750 ) gev are allowed by electroweak precision data at the @xmath0 and @xmath1 cl . within this parameter space , there are dramatic effects on higgs phenomenology : production rates are enhanced , weak boson fusion channels are suppressed , angular distributions are modified , and higgs pairs can we observed . we also identify exotic signals , such as higgs decay to same - sign dileptons . finally , we estimate the upper bound on the cutoff scale from vacuum stability and triviality .
as location - based services become widely deployed , the importance of verifying the location information being fed into the location service is becoming a critical security issue . the main difference between a location verification system ( lvs ) and a localization system is that we are confronted by some _ a priori _ information , such as a claimed position in the lvs @xcite . in the context of a main target application of our system , namely intelligent transport systems ( its ) , the issue of location verification has attracted a considerable amount of recent attention @xcite . normally , in order to infer whether a network user or node is malicious ( attempting to spoof location ) or legitimate ( actually at the claimed location ) , we have to set a threshold for the lvs . this threshold is set so as to obtain low false positive rates for legitimate users and high detection rates for malicious users . as such , the specific value of the threshold will directly affect the performance of an lvs . one traditional approach to set the threshold of an lvs is to search for a tradeoff between false positive rate and detection rate according to receiver operating characteristic ( roc ) curve @xcite . another technique is to obtain the false positive and detection rates through empirical training data and minimize specific functions of the two rates to set the threshold @xcite @xcite @xcite . for example , in @xcite , the sum of false positive and false negative rates were minimized . however , although successful in many scenarios , the approaches mentioned above do not specify in any formal sense what the ` optimal ' threshold value of an lvs should be . in addition , in our key target application of our lvs , namely its , it is not practical to collect the required training data due to the variable circumstances . the main point of this paper is to develop for the first time an information theoretic framework that will allow us to formally set the optimal threshold of an lvs . in order to do this , we first define a threshold based on the squared mahalanobis distance , which utilizes the fisher information matrix ( fim ) associated with the location information metrics utilized by the lvs . to optimize the threshold , the intrusion detection capability ( idc ) proposed by gu @xmath0 @xcite for an intrusion detection system ( ids ) will be utilized . the idc is the ratio of the reduction of uncertainty of the ids input given the output . as such , the idc measures the capability of an ids to classify the input events correctly . a larger idc means that the lvs has an improved capability of classifying users as malicious or legitimate accurately . from an information theoretic point of view the optimal threshold is the value that maximizes the idc . the rest of this paper is organized as follows . section 2 presents the system model , which details the observation model and the threat model we utilize . in section 3 , the threshold is defined in terms of the fim associated with the location metrics . section 3 also provides the techniques used to determine the false positive and detection rates , which are utilized to derive the idc . section 4 provides the details of how the idc is used in the optimizing the threshold . simulation results which validate our new analytical lvs framework are presented in section 5 . section 6 concludes and discusses some future directions . let us assume a user could obtain its true position , @xmath1 $ ] , from its localization equipment ( @xmath2 , gps ) , and that the localization error is zero . thus , a legitimate user s claimed ( reported ) position , @xmath3 $ ] , is exactly the same as its true position @xmath4 . however , a malicious user will falsify ( spoof ) its claimed position in an attempt to fool the lvs . we denote the legitimate and malicious hypothesis as @xmath5 and @xmath6 , respectively , and the _ a priori _ information can be summarized as @xmath7 although the framework we develop can be built on any location information metric , for purposes of illustration in this work we will solely investigate the case where the location information metric is the received signal strength ( rss ) obtained by a base station ( bs ) from a user . the rss of the @xmath8-th bs from a legitimate user , @xmath9 , is assumed to be given by @xmath10 where @xmath11 is a reference received power , @xmath12 is the reference distance , @xmath13 is the path loss exponent , @xmath14 is a zero - mean normal random variable with variance @xmath15 , the euclidean distance of the @xmath8-th bs to the user s true position @xmath16 $ ] is @xmath17 where @xmath18 $ ] is the location of the @xmath8-th bs , and @xmath19 is the number of bss . for @xmath5 in eq . ( [ prior ] ) , @xmath20 in eq . ( [ observ ] ) can be replaced by @xmath21 , where @xmath21 is the euclidean distance of the @xmath8-th bs to the user s claimed position @xmath22 $ ] and can be expressed as @xmath23 let us assume a malicious user knows the positions of all bss and is able to boost its transmit power according to its claimed positions . the rss of the @xmath8-th bs from a malicious user , @xmath24 , can be written as @xmath25 where @xmath26 is the boost power . we assume the malicious user is equipped with only one omni - antenna , and thus @xmath26 is constant for all the bss . in the following , one strategy to set a boost value of @xmath26 for the malicious user will be provided . a malicious user s claimed position is determined by its purpose and lvs parameters . constrained by the positions of all bss , the spoofed observations @xmath9 are not exactly the same as the ideal observations @xmath27 calculated according to its claimed position as follows @xmath28 however , the malicious user would like to spoof the observations @xmath9 as similar as possible to the ideal observations @xmath27 . thus , it will set a value of @xmath26 to minimize the divergence between @xmath24 and @xmath27 . this divergence can be defined by the mean square error ( mse ) as follows @xmath29 ^ 2\right\}\\ \nonumber & = & \frac{1}{n}\sum_{i=1}^n \left[p_x - 10\gamma \log(\frac{d_i^t}{d_0 } ) + 10\gamma \log_{10}\left(\frac{d_i^c}{d_0}\right)\right]^2,\end{aligned}\ ] ] where @xmath30 is the expectation with respect to all the observations . then , the value of @xmath26 can be expressed as @xmath31 . taking the first derivative of @xmath32 with respect to @xmath26 and setting it to zero , we can obtain @xmath33 as @xmath34 in the above we use @xmath35 instead of @xmath8 in the equations related to @xmath33 to avoid confusion them with the @xmath5 observation model . substituting @xmath33 into eq . ( [ threat ] ) , the threat model ( observation model based on @xmath6 ) can be rewritten as @xmath36 where @xmath37 eq . ( [ threat1 ] ) is the general threat model based on rss , but it is not practical since a malicious user s true position is unknown . we can approximate the threat model by assuming @xmath4 follows a distribution . here , due to the limited space , let us assume a malicious user has an approximate infinite distance away from all bss to facilitate the lvs ( the more general case is discussed later ) . given this assumption , all the bss distance s from the user converge to one value . that is , the distance of a malicious user s true position to every bs is nearly a constant number @xmath38 , @xmath2 , @xmath39 . therefore , the term @xmath40 can be rewritten as @xmath41 based on the above analysis , the threat model can be expressed as @xmath42 in this section , we first present our threshold based on the squared mahalanobis distance , which utilizes the inverse fim . then , we provide techniques used to determine the false positive rate @xmath43 and the detection rate @xmath44 of our lvs . the threshold is defined in terms of the squared mahalanobis distance of an estimated position vector @xmath45 $ ] .. in this alternative description , an error ellipse is derived directly from the fim , with the scale of the ellipse being set by @xmath46 and the orientation being set by the eigenvectors of the inverse fim . for different values of the threshold @xmath47 the ellipse size scales as @xmath48 , and the detection algorithm decides the user is malicious if the estimated position returned by the location mle lies outside of the ellipse . ] the squared mahalanobis distance can be expressed as @xcite @xmath49 where @xmath50 is the mean of @xmath51 and @xmath52 is the covariance matrix of @xmath51 . according to the definition of @xmath53 , it is a dimensionless scalar and involves not only the euclidean distance but also the geometric information . in an lvs , we are interested in the ` distance ' between a user s estimated position @xmath51 and its claimed position @xmath54 . thus , we will use @xmath54 instead of @xmath50 to calculate @xmath53 . in addition , without any _ a priori _ results from a localization algorithms , we can not obtain any estimate of the covariance matrix @xmath52 . therefore , we will utilize the inverse fim , @xmath55 , to approximate @xmath52 . with this , the squared mahalanobis distance in our lvs can be written as @xmath56 where @xmath57 and @xmath58 is the fim to be calculated as given below . in practice , the lvs works on the observation model based on @xmath5 , and the likelihood function of received powers can be obtained using eq . ( [ observ ] ) . let us assume the observations received by different bss are independent , then the log - likelihood function can be expressed as @xmath59 ^ 2 + \log \mathbf{c}.\end{aligned}\ ] ] where @xmath60 is the @xmath19-dimension observation vector and the constant number @xmath61 is @xmath62 then , we can calculate the terms of the fim through @xmath63,\end{aligned}\ ] ] where @xmath30 represents the expectation operation with respect to all observations . after some algebra , the fim can be written as @xcite , @xmath64,\end{aligned}\ ] ] where @xmath65 after setting a threshold parameter @xmath47 for the squared mahalanobis distance , the decision rule of an lvs ( _ i.e. _ a malicious user or not ) can be expressed as follows @xmath66 note that , we are able to transform any covariance matrix into a diagonal matrix by rotating the position vector @xcite . thus , the general form of @xmath55 can be expressed as @xmath67}.\end{aligned}\ ] ] then , the threshold @xmath47 can be encapsulated within the equation for an ellipse as follows @xmath68 therefore , the threshold @xmath47 can also be understood as an ellipse , denoted as @xmath69 , which is determined by extending the error ellipse provided by the fim with the threshold parameter @xmath47 . based on the above analysis , the overall process of an lvs includes four steps * collect observations of the rss received from a user by each bs ; * apply a localization algorithm to obtain an estimated position @xmath70 ; * calculate the squared mahalanobis distance @xmath71 of @xmath70 to the user s claimed position @xmath72 ; * infer if the user is legitimate or malicious according to the decision rule in eq . ( [ rule ] ) . in practice , the above are all the steps of our lvs . however , to evaluate an lvs , false positive and detection rates , which are functions of the threshold parameter @xmath47 and other lvs parameters , are always investigated in theory . in the following subsections , we provide techniques used to determine false positive and detection rates in order to optimize the threshold parameter @xmath47 . the false positive rate @xmath43 is the probability by which legitimate users are judged as malicious ones . for a legitimate user , @xmath73 . then , in the 2-d physical space , the false positive rate can be expressed as @xmath74 @xcite . in fact , the true positive rate ( @xmath75 ) is a well known metric that underlies the performance of unbiased localization algorithms . for example , in the 2-d physical space , it states that the probability by which an estimated position lies within the ellipse with @xmath76 is no more than @xmath77 . the detection rate @xmath44 is the probability that malicious users are recognized as malicious ones . in order to calculate @xmath44 , we have to obtain the posterior probability density function ( pdf ) for a location given some rss observation vector , which can be expressed as @xmath78 where @xmath79 $ ] is a general location , and @xmath80 $ ] is the observation vector . of course , if the user is malicious the observed signal vector @xmath60 will be one that has undergone a boost as described by eq . ( [ threat2 ] ) . let us denote the average value of this spoofed observation vector as @xmath81 . given this , the likelihood function @xmath82 can be derived from eq . ( [ observ ] ) . if we take @xmath83 to be a uniform variable vector , then the detection rate @xmath44 can be calculated as @xmath84\in \mathbb{t}}}f(\bm{\theta}\|\bm{\hat{p}})dx dy = 1-\frac{1}{a_1 } { \mathop{\int\int}\limits_{[x , y]\in \mathbb{t}}}f(\bm{\hat{p}}\|\bm{\theta})dx dy,\end{aligned}\ ] ] where @xmath85 is a normalizing constant that can be written as @xmath86 where @xmath87 numerical methods are utilized to solve the above integral equation for @xmath44 since there is no closed form solution . based on the above analysis , @xmath44 is also a function of @xmath47 . as an aside it is worth mentioning that the false positive rate @xmath43 can also be written in a similar form as follows @xmath88\in \mathbb{t}}}f(\bm{\tilde{p}}\|\bm{\theta})dx dy,\end{aligned}\ ] ] where @xmath89 is the average non - spoofed observation vector and @xmath90 where @xmath91 in this section we will optimize the value of the threshold by maximizing the idc , which is a function of the false positive rate @xmath43 , detection rate @xmath44 and the base rate @xmath92 ( the _ a priori _ probability of intrusion in the input event data ) . that is , our optimization procedure is to find the value of @xmath47 that maximizes the idc . from an information theoretic point of view , the idc is a metric that measures the capability of an ids to classify the input events correctly and is defined as @xcite @xmath93 where @xmath94 is the entropy of the input data @xmath95 , @xmath96 is the mutual information of input data @xmath95 and output data @xmath97 , and @xmath98 is the conditional entropy . mutual information @xmath99 measures the reduction of uncertainty of the input @xmath100 given the output @xmath101 . thus , @xmath102 is the ratio of the reduction of uncertainty of the input given the output . its value range is [ 0 , 1 ] . a larger @xmath102 value means that the ids has an improved capability of classifying input events accurately . our lvs can be modeled as an ids whose input data are the claimed positions , and the output data are the binary decisions . then , @xmath103 represents an actual claimed position from a legitimate user , @xmath104 represents a spoofed claimed position from a malicious user , @xmath105 infers the user is legitimate , and @xmath106 indicates the user is malicious . accordingly , the false positive rate @xmath43 is the probability @xmath107 , and detection rate @xmath44 is the probability @xmath108 . therefore , the optimal value of @xmath47 is the one that maximizes the value of the @xmath102 of the lvs . the realizations of input and output data are denoted as @xmath109 and @xmath110 , respectively . given the base rate @xmath92 , the entropy of the input data @xmath94 can be written as @xcite @xmath111 the conditional entropy @xmath98 can be expressed as @xmath112 numerical methods are applied in order to search for the optimal value of @xmath47 since there is no closed form for @xmath44 . in the following we refer to this optimal value as @xmath113 . adopting a maximum likelihood estimator ( mle ) in our location estimation algorithm we now verify , via detailed simulations , our previous analysis . the theoretical and simulated @xmath43 , @xmath44 and @xmath102 , all of which are dependent on @xmath47 , are utilized in order to find the value @xmath113 that maximizes @xmath102 . the simulation settings are as follows : * @xmath19 bss are deployed in a @xmath114 square field and the legitimate and honest users can communicate with all bss ; * the claimed positions of honest and malicious users are the same , denoted @xmath72 ; * @xmath115 observations are collected from each base station ; * the bss are set at fixed positions ( we investigate a range of fixed locations ) ; * the results shown are averaged over 1,000 monte carlo realizations of the estimated position , and where the base rate @xmath116 for all the simulations . as shown in fig.1 , the solid lines are the theoretical @xmath43 , @xmath44 and @xmath102 while the symbols are the simulated @xmath43 , @xmath44 and @xmath102 . the simulated values of @xmath43 and @xmath44 are calculated directly according to the realizations of estimated positions , and then the simulated @xmath102 is obtained from eq . ( [ h7 ] ) . the simulation parameters are shown in the figure caption and the theoretical optimal value @xmath113 can be seen to be @xmath118 ( note that in all the figures explicitly shown in this paper the four bss are fixed at the corners of a 200 m x 200 m grid ) . the comparison between simulation and analysis shows excellent agreement . beyond the simulations explicitly shown in fig.1 , we have investigated a range of other fixed bss positions ( up to 10 bss whose positions are randomly selected ) , and these simulation also show excellent agreement with simulations . collectively , these simulation results verify the analysis we have provided earlier . the simulation results with a malicious user having a certain distance to all bss are shown in fig.2 . the true position of the malicious user in the simulations is set at 10 km away from the claimed position . although the simulation and theoretical values of @xmath43 , @xmath44 and @xmath102 do not match with each other exactly ( the theoretical analysis approximates the user as being at infinity ) , the simulation and theoretical optimal values @xmath113 are effectively the same . we find this result holds down to distance where the malicious user is a few km away from the claimed position . this shows that our framework is tenable when the assumption that malicious user is infinitely far away is relaxed down to the few km range . in order to verify the @xmath102 with the optimal value @xmath113 is correct , we also simulated @xmath102 for a range of @xmath15 . 3 shows such results for the case where the user malicious user if effectively at infinity . here the optimal value @xmath113 is derived from the proposed theoretical analysis , but in the simulations the threshold is set to the other values of t shown ( @xmath119 and @xmath120 ) . from the results shown we can see that these other values do provide simulated false positive and detection rates which result in lower values of @xmath102 ( and therefore sub - optimal performance ) , which once again verifies the robustness of our analytical framework . 4 shows the same results except that the malicious user is again set at 10 km away from the claimed position . again we see a validation of our analysis . in this paper , we have proposed a novel and rigorous information theoretic framework for an lvs . the theoretical framework we have developed shows how the value of the threshold used in the detection of a spoofed location can be optimized in terms of the mutual information between the input and output data . in order to verify the legitimacy of our framework we have carried out detailed numerical simulations of our framework under the assumption of an idealized threat model in which the malicious user is far enough from the claimed location such that his boosted signal strength results in all bss receiving the same rss ( modulo noise ) . our numerical simulations mimic the practical scenario where a system deployed using our framework must make a binary yes / no malicious decision " to each snapshot of rss values obtained by the bss . the comparison between simulation and analysis shows excellent agreement . other simulations where we modify the approximation of constant rss at bss also showed very good agreement with analysis . the work described in this paper formalises the performance of an optimal lvs system under the simplest ( and perhaps most likely scenario ) , where a single malicious user attempts to spoof his location to a wider wireless network . the practical scenario we had in mind whilst carrying out our simulations was in an its where another vehicle is attempting to provide falsified location information the wider vehicular network . future work related our new framework will include the formal inclusion of more sophisticated threat models , where the malicious user is both closer to the claimed location and has the use of colluding adversaries . it is well known that no lvs can be made foolproof under the colluding adversary scenario , however , we will investigate in a formal information theoretic sense the detailed nature of the vulnerability of an lvs under such different threat models . this work has been supported by the university of new south wales , and the australian research council ( arc ) . s. @xmath121apkun , k. b. rasmussen , m. @xmath121agalj , and m. srivastava , secure location verification with hidden and mobile base station , " _ ieee trans . mobile comput . _ , vol . 7 , no . 470 - 483 , apr . 2008 . h. buhrman , n. chandran , s. fehr , r. gelles , v. goyal , r. ostrovsky and c. schaffner , position - based quantum cryptography : impossibility and constructions , " in advances in cryptology , vol . 6841 of lecture notes in computer science , pp . 429 - 446 , springer - verlag , 2011 .	as location - based applications become ubiquitous in emerging wireless networks , location verification systems ( lvs ) are of growing importance . in this paper we propose , for the first time , a rigorous information - theoretic framework for an lvs . the theoretical framework we develop illustrates how the threshold used in the detection of a spoofed location can be optimized in terms of the mutual information between the input and output data of the lvs . in order to verify the legitimacy of our analytical framework we have carried out detailed numerical simulations . our simulations mimic the practical scenario where a system deployed using our framework must make a binary yes / no malicious decision " to each snapshot of the signal strength values obtained by base stations . the comparison between simulation and analysis shows excellent agreement . our optimized lvs framework provides a defence against location spoofing attacks in emerging wireless networks such as those envisioned for intelligent transport systems , where verification of location information is of paramount importance .
the modified newtonian dynamics ( mond ) @xcite is a non - linear theory of gravity which undergoes departures from the standard newtonian inverse - square law at a characteristic acceleration scale ; it was proposed to explain certain features of the motion of ordinary electromagnetically detectable matter in galaxies and of galaxies in galactic clusters without resorting to exotic forms of still undetected dark matter . mond , among other things , does not fulfill the strong equivalence principle in the sense that the internal dynamics of a gravitating system @xmath0 of bodies does depend on the external background gravitational field @xmath2 of a larger system @xmath1 in which @xmath0 is embedded , even if @xmath2 is constant and uniform @xcite ; it is the so - called external field effect ( efe ) . its effects have mainly been treated in a number of ways in the context of galactic dynamics @xcite , while they have received some attention in the case of the oort cloud in the solar system @xcite . it is just the case to recall that it moves through the milky way at about 8.5 kpc from its center ; its gravitational attraction at the solar system s location can be evaluated from the magnitude of its centrifugal acceleration @xcite @xmath15 , where @xcite @xmath16 m s@xmath5 is the mond characteristic acceleration scale . thus , in principle , such an efe should affect the inner dynamics of the solar system . concerning such an issue , a researcher active in mond writes : he / she also adds that another researcher working on mond tries to go in deeper details by writing : such statements are likely expressions of a widely diffuse belief about efe in solar system . in this paper we deal with statements like those previously quoted by showing , in a purely phenomenological way , that they are , in fact , not valid for the motion of the major bodies of the solar system referred to the solar system barycenter ( ssb ) frame usually adopted for planetary data reduction . in particular , we will address the following points * does a constant and uniform acceleration @xmath2 having a generic direction @xmath17 and a magnitude of the order of @xmath18 m s@xmath5 affect the motion of the major bodies of the solar system ? * are the effects of such an acceleration negligible for the sun s planets , although they are , in principle , present ? * since @xmath2 is constant and uniform , are its effects canceled out over planetary revolutions ? we will demonstrate that * a constant and uniform acceleration does induce non - zero long - period , i.e. averaged over one orbital revolution , effects on the keplerian orbital elements of a planet * by assuming @xmath19 , the resulting perihelion precessions of the inner planets are @xmath9 orders of magnitude larger than the present - day limits on the recently estimated non - standard perihelion rates @xcite . such an external field @xmath2 , which would not be aligned with the internal newtonian attraction @xmath20 , but it would be directed along a generic direction @xmath17 fixed during one orbital revolution , can be modelled as e = e_x i + e_y j + e_z k , where @xmath21 are assumed constant and uniform , and @xmath22 are the unit vectors of the usual ssb frame used to describe the planetary motions . the constraints we will obtain are , thus , of dynamical origin and model - independent . from an observational point of view , the astronomer e. v. @xcite has recently estimated , in a least - square sense , corrections @xmath10 to the standard newtonian / einsteinian averaged precessions of the longitudes of the perihelia of the inner planets of the solar system , shown in table [ chebolas ] , .estimated corrections @xmath10 , in units of @xmath23 arcsec cty@xmath24 ( 1 arcsec cty@xmath25 s@xmath24 ) , to the standard newton / einstein perihelion precessions of the inner planets according to table 3 of @xcite ( mercury , earth , mars ) . the result for venus has been obtained by recently processing radiometric data from magellan spacecraft ( e.v . pitjeva , private communication , 2008 ) . the reference frame used is the usual solar system barycentric one.[chebolas ] [ cols="^,^,^,^ " , ] as a result , we have for the cartesian components of the perturbing acceleration @xmath2 the figures quoted in table [ acci ] . we have obtained them as follows . first , we have considered the eight different systems of the type of obtained by taking the maximum and the minimum values of @xmath10 for venus , earth and mars according to their best estimates and uncertainties of table [ chebolas ] ; then , after solving them , we computed the means and the standard deviations of the three sets of eight solutions for @xmath21 . the upper bounds on @xmath21 are @xmath11 m s@xmath5 , @xmath26 m s@xmath5 , @xmath13 m s@xmath5 , respectively . the total acceleration is e = ( 0.63)10 ^ -14 m s^-2.such a constrain , which is a conservative one since we linearly added the uncertainties in @xmath21 to obtain e , is four orders of magnitude smaller than @xmath27 . this substantially confirms our preliminary , non - analytical evaluation . how could we interpret the results obtained here ? in fact , there is no contradiction with the findings in @xcite because they are valid in an inertial galactocentric frame , for which the boundary condition at infinity yielding -e for r , where @xmath28 is the solution of the modified poisson equation for the gravitational potential in the non - relativistic theory of mond proposed by @xcite , holds . see figure [ gcframe ] in which we have indicated the newtonian monopole term with @xmath20 and the galactic attraction @xmath2 with @xmath29 . sun - planet system @xmath0 embedded in an external system @xmath1 given by the milky way : galactocentric frame . in it @xmath20 ( red arrow ) is the newtonian gravitational attraction of the planet by the sun ( the newtonian gravitational attraction of the sun by the planet is not depicted ) , while @xmath30 ( light green arrow ) is the gravitational attraction of the planet by the galaxy whose magnitude amounts to about @xmath8 m s@xmath5 ( in principle , it should also account for the gravitational attraction by other sources external to the galaxy itself as the galaxy m31 andromeda , etc . : however , their acceleration is of the order of @xmath31 ) . ] it is not so in the ssb frame used for usual solar system s planetary data reduction because of the centrifugal acceleration which ( almost ) exactly cancels the galactic gravitational attraction throughout ssb leaving just a small net tidal residual @xmath32 , if any , whose magnitude can be easily evaluated as @xmath33 m s@xmath5 . see figure [ ssbframe ] . sun - planet system @xmath0 embedded in an external system @xmath1 given by the milky way : heliocentric frame . in it @xmath20 is the newtonian gravitational attraction of the planet by the sun ( red arrow ) , while @xmath2 is given by the gravitational tidal forces @xmath32 ( olive green arrow ) exerted on the planet by the galaxy whose magnitude amounts to about @xmath33 m s@xmath5 ; they account for both the uniform centrifugal accelerations due to the motion of solar system through the galaxy ( light blue arrow ) and the non - uniform gravitational attraction by the galaxy throughout the extension of the solar system ( light green arrow ) . ] in other words , the outcome of the gravitational experiment represented by the planetary motion does depend on the velocities of the inertial frames in which it is studied , which are gc ( at rest ) and ssb ( freely falling in the milky way ) . after all , this should not surprise too much , since mond does , indeed , violate the strong equivalence principle . in this paper we dealt with the problem of the external field effect in solar system in the framework of the modified newtonian dynamics . in particular , we analyzed the common belief that , in principle , also the internal motions of the sun s planets are affected , in mond , by a constant and uniform acceleration @xmath2 having the same magnitude of the centrifugal acceleration of the solar system s revolution around the galactic center @xmath4 m s@xmath5 , and directed along a fixed direction @xmath17 in space . it is believed that its magnitude is too small to affect the motions of the major bodies of the solar system , and that its net effects vanish because of its constancy during one planetary orbital revolution . we found , instead , that such a kind of perturbing acceleration does , in fact , produce non - vanishing net long - period , i.e. averaged over one orbital revolution , orbital effects on the motion of a test particle . among them , of particular relevance are the perihelion precessions because it is possible to compare them to the corrections to the standard newtonian / einsteinian perihelion rates of the inner planets recently estimated with the epm ephemerides in the ssb frame . as a result , we obtained upper bounds on the components of @xmath2 by finding @xmath11 m s@xmath5 , @xmath12 m s@xmath5 , @xmath13 m s@xmath5 , with @xmath34 m s@xmath5 , i.e. @xmath9 orders of magnitude smaller than @xmath35 . this result can be interpreted in terms of a violation of the independence of the outcome of a gravitational experiment from the velocity of the frame in which it is performed , in according with the fact that mond violates the strong equivalence principle . our results are valid not only for efe in mond , but also for any other putative exotic modification of gravity yielding a constant and uniform net small extra - acceleration which , thus , can be no larger than @xmath36 m s@xmath5 in the solar system . it will be important to repeat such a test if and when other teams of astronomers will independently estimate their own corrections to the standard perihelion precessions with different ephemerides . pitjeva e.v . , 2008 , use of optical and radio astrometric observations of planets , satellites and spacecraft for ephemeris astronomy , in a giant step : from milli- to micro - arcsecond astrometry proceedings iau symposium no . 248 , 2007 ed . w. j. jin , i. platais & m. a. c. perryman ( cambridge : cambridge university press ) , 20	in the framework of the modified newtonian dynamics ( mond ) , the internal dynamics of a gravitating system @xmath0 embedded in a larger one @xmath1 is affected by the external background field @xmath2 of @xmath1 even if it is constant and uniform , thus implying a violation of the strong equivalence principle : it is the so - called external field effect ( efe ) . in the case of the solar system , @xmath3 would be @xmath4 m s@xmath5 because of its motion through the milky way : it is orders of magnitude smaller than the main newtonian monopole terms for the planets . we address here the following questions in a purely phenomenological manner : are the sun s planets affected by an efe as large as @xmath6 m s@xmath5 ? can it be assumed that its effect is negligible for them because of its relatively small size ? does @xmath2 induce vanishing net orbital effects because of its constancy over typical solar system s planetary orbital periods ? it turns out that a constant and uniform acceleration , treated perturbatively , does induce non - vanishing long - period orbital effects on the longitude of the pericenter @xmath7 of a test particle . in the case of the inner planets of the solar system and with @xmath8 m s@xmath5 , they are @xmath9 orders of magnitude larger than the present - day upper bounds on the non - standard perihelion precessions @xmath10 recently obtained with by e.v . pitjeva with the epm ephemerides in the solar system barycentric frame . the upper limits on the components of @xmath2 are @xmath11 m s@xmath5 , @xmath12 m s@xmath5 , @xmath13 m s@xmath5 . this result is in agreement with the violation of the strong equivalence principle by mond . our analysis also holds for any other exotic modification of the current laws of gravity yielding a constant and uniform extra - acceleration . if and when other corrections @xmath10 to the usual perihelion precessions will be independently estimated with different ephemerides it will be possible to repeat such a test . # 1 # 1 # 1#2_#1^#2 # 1eq . ( [ # 1 ] ) # 1#2eq . ( [ # 1])-eq . ( [ # 2 ] ) # 1eq . ( [ # 1 ] ) # 1#2eq . ( [ # 1])-eq . ( [ # 2 ] ) # 1 # 1``#1 '' # 1#2#3 # 1#2 # 1#2 # 1#2 # 1ref . # 1 # 1#2 # 1#2 # 1#2 # 1 \|@xmath14 # 1#2 # 1 # 1[#1 ] # 1#1 # 1(#1 ) # 1
the work is supported in part by the national science council of r.o.c . under grant # s : nsc-95 - 2112-m-007 - 059-my3 and nsc-98 - 2112-m-007 - 008-my3 and national tsing hua university under the boost program and grant # : 99n2539e1 . for a review , see e. j. copeland , m. sami and s. tsujikawa , int . j. mod . d * 15 * , 1753 ( 2006 ) . for recent reviews , see t. p. sotiriou and v. faraoni , rev . phys . * 82 * , 451 ( 2010 ) ; a. de felice and s. tsujikawa , arxiv:1002.4928 [ gr - qc ] . s. nojiri and s. d. odintsov , phys . d * 68 * , 123512 ( 2003 ) ; a. d. dolgov and m. kawasaki , phys . b * 573 * , 1 ( 2003 ) ; v. faraoni , phys . rev . d * 74 * , 104017 ( 2006 ) ; y. s. song , w. hu and i. sawicki , _ ibid_. * 75 * , 044004 ( 2007 ) . v. muller , h. j. schmidt and a. a. starobinsky , phys . b * 202 * , 198 ( 1988 ) ; v. faraoni and s. nadeau , phys . d * 72 * , 124005 ( 2005 ) ; l. amendola , r. gannouji , d. polarski and s. tsujikawa , _ ibid_. * 75 * , 083504 ( 2007 ) . t. chiba , phys . b * 575 * , 1 ( 2003 ) ; t. chiba , t. l. smith and a. l. erickcek , phys . d * 75 * , 124014 ( 2007 ) . w. hu and i. sawicki , phys . d * 76 * , 064004 ( 2007 ) . a. a. starobinsky , jetp lett . * 86 * , 157 ( 2007 ) . s. tsujikawa , phys . d * 77 * , 023507 ( 2008 ) . p. zhang , phys . d * 73 * , 123504 ( 2006 ) ; b. li and m. c. chu , _ ibid_. * 74 * , 104010 ( 2006 ) ; r. bean , d. bernat , l. pogosian , a. silvestri and m. trodden , _ ibid_. * 75 * , 064020 ( 2007 ) ; p. j. zhang , _ ibid_. * 76 * , 024007 ( 2007 ) ; p. zhang , m. liguori , r. bean and s. dodelson , phys . rev . lett . * 99 * , 141302 ( 2007 ) ; g. calcagni and g. nardelli , arxiv:1004.5144 [ hep - th ] . g. cognola , e. elizalde , s. nojiri , s. d. odintsov , l. sebastiani and s. zerbini , phys . d * 77 * , 046009 ( 2008 ) . e. v. linder , phys . d * 80 * , 123528 ( 2009 ) . k. bamba , c. q. geng and c. c. lee , jcap * 1008 * , 021 ( 2010 ) [ arxiv:1005.4574 [ astro-ph.co ] ] . k. bamba , c. q. geng and c. c. lee , jcap * 1011 * , 001 ( 2010 ) [ arxiv:1007.0482 [ astro-ph.co ] ] . u. alam , v. sahni and a. a. starobinsky , jcap * 0406 * , 008 ( 2004 ) ; * 0702 * , 011 ( 2007 ) ; s. nesseris and l. perivolaropoulos , _ ibid_. * 0701 * , 018 ( 2007 ) ; p. u. wu and h. w. yu , phys . b * 643 * , 315 ( 2006 ) ; h. k. jassal , j. s. bagla and t. padmanabhan , arxiv : astro - ph/0601389 . r. r. caldwell , phys . b * 545 * , 23 ( 2002 ) ; b. feng , x. l. wang and x. m. zhang , _ ibid_. * 607 * , 35 ( 2005 ) ; z. k. guo , y. s. piao , x. m. zhang and y. z. zhang , _ ibid_. * 608 * , 177 ( 2005 ) . m. martinelli , a. melchiorri and l. amendola , phys . d * 79 * , 123516 ( 2009 ) ; h. motohashi , a. a. starobinsky and j. yokoyama , arxiv:1002.1141 [ astro-ph.co ] ; arxiv:1005.1171 [ astro-ph.co ] . m. c. b. abdalla , s. nojiri and s. d. odintsov , class . quant . grav . * 22 * , l35 ( 2005 ) ; i. h. brevik , int . j. mod . d * 15 * , 767 ( 2006 ) ; gen . grav . * 38 * , 1317 ( 2006 ) ; l. amendola and s. tsujikawa , phys . lett . b * 660 * , 125 ( 2008 ) ; k. bamba , c. q. geng , s. nojiri and s. d. odintsov , phys . d * 79 * , 083014 ( 2009 ) ; mod . a * 25 * , 900 ( 2010 ) ; k. bamba and c. q. geng , phys . b * 679 * , 282 ( 2009 ) ; k. bamba , arxiv:0904.2655 [ gr - qc ] ; k. bamba , s. nojiri and s. d. odintsov , jcap * 0810 * , 045 ( 2008 ) ; y. bisabr , arxiv:1005.5679 [ gr - qc ] . k. bamba and c. q. geng , prog . phys . * 122 * , 1267 ( 2009 ) . s. capozziello and s. tsujikawa , phys . d * 77 * , 107501 ( 2008 ) ; a. ali , r. gannouji , m. sami and a. a. sen , _ ibid_. * 81 * , 104029 ( 2010 ) . s. tsujikawa , r. gannouji , b. moraes and d. polarski , phys . rev . d * 80 * , 084044 ( 2009 ) . s. a. appleby and r. a. battye , phys . b * 654 * , 7 ( 2007 ) ; s. appleby , r. battye and a. starobinsky , jcap * 1006 * , 005 ( 2010 ) .	we review the equation of state for dark energy in modified gravity theories . in particular , we summarize the generic feature of the phantom divide crossing in the past and future in viable @xmath0 gravity models . to understand the late time acceleration universe @xcite , one of the interesting possibilities is to consider a modified gravitational theory , such as @xmath0 gravity @xcite . to build up a viable @xmath0 gravity model , one needs to satisfy the following conditions : ( a ) positivity of the effective gravitational coupling , ( b ) stability of cosmological perturbations @xcite , ( c ) asymptotic behavior to the standard @xmath1-cold - dark - matter ( @xmath2 ) model in the large curvature regime , ( d ) stability of the late - time de sitter point @xcite , ( e ) constraints from the equivalence principle , and ( f ) solar - system constraints @xcite . several viable models have been constructed in the literature , such as the following popular ones:@xcite @ccccc@ + & & model & @xmath0 & constant parameters + + & & ( i ) hu - sawicki & @xmath3 & @xmath4 , @xmath5 , @xmath6,@xmath7 + & & ( ii ) starobinsky & @xmath8 $ ] & @xmath9 , @xmath10 , @xmath11 + & & ( iii ) tsujikawa&@xmath12 & @xmath13 , @xmath14 + & & ( iv ) exponential & @xmath15 & @xmath16 , @xmath17 + + [ table ] recently , the cosmological observational data @xcite seems to indicate the crossing of the phantom divide @xmath18 of the equation of state for dark energy in the near past . to understand such a crossing , many attempts have been made . the most noticeable one is to use a phantom field with a negative kinetic energy term @xcite . clearly , it surfers a serious problem as it is not stable at the quantum level . on the other hand , the crossing of the phantom divide can also be realized in the above viable @xmath0 models @xcite without violating any stability conditions . this is probably the most peculiar character of the modified gravitational models . other @xmath0 gravity models with realizing a crossing @xcite as well as multiple crossings @xcite of the phantom boundary have also been examined . in this talk , we would like to review equation of state in @xmath0 gravity . in particular , we show that the viable @xmath0 models generally exhibit the crossings of the phantom divide in the past as well as future @xcite . the action of @xmath0 gravity with matter is given by @xmath19 where @xmath20 is the determinant of the metric tensor @xmath21 , @xmath22 is the action of matter which is assumed to be minimally coupled to gravity , i.e. , the action @xmath23 is written in the jordan frame , and @xmath24 denotes matter fields . here , we use the standard metric formalism . by taking the variation of the action in eq . ( [ eq:1 ] ) with respect to @xmath21 , one obtains @xcite @xmath25 where @xmath26 is the einstein tensor , @xmath27 , @xmath28 is the covariant derivative operator associated with @xmath29 , @xmath30 is the covariant dalembertian for a scalar field , and @xmath31 is the contribution to the energy - momentum tensor from all perfect fluids of matter . from eq . ( [ eq:2 ] ) , we obtain the following gravitational field equations : @xmath32 where @xmath33 is the hubble parameter , the dot denotes the time derivative of @xmath34 , and @xmath35 and @xmath36 are the energy density and pressure of all perfect fluids of matter , respectively . the equation of state for dark energy is given by @xmath37 where @xmath38\ , , \nonumber\\ % \end{equation } % where % \begin{eqnarray } p_{\mathrm{de } } & = & \frac{1}{\kappa^2 } \left [ -\frac{1}{2 } \left ( fr - f \right ) + \ddot{f}+2h \dot{f } -\left(1-f\right)\left(2\dot{h}+3h^2\right ) \right]\ , . \label{eq : w}\end{aligned}\ ] ] + ll + ll + ll + in figs . 1 , 2 and 3 , we depict the evolution of @xmath39 , future evolutions of @xmath40 , and @xmath41 with @xmath42 and @xmath43 , as functions of the redshift @xmath44 in ( i ) hu - sawicki model for @xmath45 , @xmath46 and @xmath47 , ( ii ) starobinsky model for @xmath48 and @xmath49 , ( iii ) tsujikawa model for @xmath50 and ( iv ) the exponential gravity model for @xmath51 , respectively , and the subscript ` f ' denotes the value at the final stage @xmath52 . note that the present time is @xmath53 and the future is @xmath54 . the parameters used for each model in figs . 13 are the viable ones @xcite . several remarks are as follows : ( a ) the qualitative results do not strongly depend on the values of the parameters in each model ; ( b ) we have studied the appleby - battye model @xcite , which is also a viable @xmath0 model , and we have found that the numerical results are similar to those in the starobinsky model of ( ii ) as expected . we note that the present values of @xmath55 are -0.92 , -0.97 , -0.92 and -0.93 for the models of ( i)(iv ) , respectively . these values satisfy the present observational constraints @xcite . moreover , a dimensionless quantity @xmath56 can be determined through the numerical calculations , where @xmath57 is the energy density of non - relativistic matter at the present time . if we use the observational data on the current density parameter of non - relativistic matter @xmath58 with @xmath59 @xcite , we find that the present value of the hubble parameter @xmath60 is @xmath61 @xcite for all the models of ( i)(iv ) . furthermore , @xmath62 = 0.80 , 0.85 , 0.78 and 0.81 for the models of ( i)(iv ) , respectively , where @xmath43 . it is clear from figs . 13 that in the future ( @xmath63 ) , the crossings of the phantom divide are the generic feature for all the existing viable @xmath0 models . by writing the first future crossing of the phantom divide and the first sign change of @xmath64 from negative to positive as @xmath65 and @xmath66 , respectively , we find that @xmath67 , @xmath68 , @xmath69 and @xmath70 , where the subscript @xmath71 represents the @xmath71th viable model . the values of the ratio @xmath72 at @xmath65 and @xmath66 are @xmath73 , @xmath74 , @xmath75 and @xmath76 , where @xmath77 and @xmath78 are the density parameters of dark energy and non - relativistic matter ( cold dark matter and baryon ) , respectively . as @xmath79 decreases ( @xmath80 ) , dark energy becomes much more dominant over non - relativistic matter ( @xmath81 ) . as a result , one has @xmath82 , where @xmath83 is the effective equation of state for the universe , and @xmath84 and @xmath85 are the total energy density and pressure of the universe , respectively . here , @xmath86 and @xmath87 are the energy density of non - relativistic matter ( radiation ) and the pressure of radiation , respectively . the physical reason why the crossing of the phantom divide appears in the farther future ( @xmath80 ) is that the sign of @xmath64 changes from negative to positive due to the dominance of dark energy over non - relativistic matter . as @xmath88 in the farther future , @xmath39 oscillates around the phantom divide line @xmath18 because the sign of @xmath64 changes and consequently multiple crossings can be realized . finally , we mention that in our numerical calculations , we have taken the initial conditions of @xmath89 , @xmath90 , @xmath91 and @xmath92 for the models of ( i)(iv ) at @xmath93 , respectively , so that @xmath94 with @xmath95 , to ensure that they can be all close enough to the @xmath2 model with @xmath96 . in this talk , we have explored the past and future evolutions of @xmath39 in the viable @xmath0 gravity models and explicitly shown that the crossings of the phantom divide are the generic feature in these models . we have demonstrated that in the future the sign of @xmath64 changes from negative to positive due to the dominance of dark energy over non - relativistic matter . this is a common physical phenomena to the existing viable @xmath0 models and thus it is one of the peculiar properties of @xmath0 gravity models characterizing the deviation from the @xmath2 model .
over the last two decades , interest in ceria based materials has grown steadily . the facile redox conversion in combination with the fact that retains its fluorite crystal structure even with significant degrees of reduction makes ceria based materials ideal for controlling the oxygen atmosphere in three - way catalysts.@xcite furthermore , the introduction of dopants like zr has been shown to increase the oxygen storage capacity , increasing the range of practical applicability.@xcite the remarkable redox properties of make ceria based materials also of interest in several other industrially important applications , such as : catalytic support , thermal barrier coatings , ionic conductors and fuel cells.@xcite more recently , ceo@xmath3 and ceria based materials have been used as thin film buffer layers in coated superconductors.@xcite + for all these applications there is a constant search for ways to improve the properties of ceria based materials even further . one of the most common ways for such improvements is through the introduction of dopants , with regimes ranging from light doping ( @xmath4% ) to mixed oxides containing @xmath5% dopants or more.@xcite different applications also require different properties to be optimized . unfortunately , optimization towards one application may worsen the performance of the system for another . as an example , for ionic conductors one wishes to maximize the ionic conductivity , while for a buffer layer one often wishes to minimize this quantity . such opposing and varying requirements make a more general theoretical study of the influence of dopants on the properties of of general interest . this is the particular aim of the present paper . + experimentally , lanthanide elements are most often studied as dopants , in addition to standard catalyst @xmath1-block metals ( au , pd , cu , ... ) .@xcite since all these elements can act as aliovalent dopants , charge compensating oxygen vacancies are introduced , contributing to the changes in the properties . to have a clear understanding of the influence dopants have on the properties of , it is important to separate the contributions of these two structural changes ; _ i.e. _ to distinguish between the consequences of doping and the subsequently introduced charge compensating vacancies . theoretical calculations are ideally suited to separate the different contributions . due to the tetravalent nature of ce in , group iv elements are the perfect starting point for such investigations . + only few studies addressed group iv dopants in ceria based materials , although for example silica is a widely used catalyst support.@xcite this limited experimental work originates from the difficulty of forming solid solutions with in a controlled way . improved redox properties compared to pure and zr doped have been presented for mixed - and - materials.@xcite andersson _ et al . _ studied the redox properties of tetravalent dopants at low dopant concentrations of @xmath6% and found a correlation between the reducibility and the dopant s ionic radius.@xcite tang _ _ investigated the redox properties of tetravalent dopants ( mn , pr , sn and zr ) and linked the lower oxygen vacancy formation energy to structural distortions and electronic modifications.@xcite yashima studied tetragonal crystal phases of and found a tetragonal phase to appear for zr concentrations @xmath7% , and a cubic phase with displaced o atoms at lower concentrations.@xcite we previously found vegard law behavior for group iv doped and proposed the shannon crystal radius instead of the ionic radius as parameter for the lattice expansion.@xcite + in the present paper , we investigate the modification of the mechanical properties of due to doping with group iv elements using density functional theory ( dft ) . we show there is a difference in energetics for doping between group iva and ivb elements , and link this to the different valence electronic structure . [ ceo2p4:ssc_resultsef][ceo2p4:ssc_resultsdos ] ) . an inverse relation between the bulk modulus and the thermal expansion coefficients is presented ( sec . [ ceo2p4:ssc_resultsbmtec ] ) and the change in the charge distribution around the dopants discussed ( sec . [ ceo2p4:ssc_resultshi ] ) . technical details of the computational methods used are provided in appendix [ ceo2p4:sc_theormeth ] . l\|cccccc & structure & pearson symbol & space group + & & & + + c & diamond & cf@xmath8 & fd@xmath9 m + si & diamond & cf@xmath8 & fd@xmath9 m + ge & diamond & cf@xmath8 & fd@xmath9 m + sn & diamond & cf@xmath8 & fd@xmath9 m + pb & face centered cubic & cf@xmath10 & fm@xmath9 m + + ti & hexagonal close packed & hp@xmath11 & p@xmath12/mmc + zr & hexagonal close packed & hp@xmath11 & p@xmath12/mmc + hf & hexagonal close packed & hp@xmath11 & p@xmath12/mmc + + co@xmath3 & dry ice & cp@xmath13 & pa@xmath9 + sio@xmath3 & @xmath14-quartz & cf@xmath8 & p@xmath15 + geo@xmath3 & rutile ( argutite ) & tp@xmath16 & p@xmath17/mnm + sno@xmath3 & rutile ( cassiterite ) & tp@xmath16 & p@xmath17/mnm + pbo@xmath3 & rutile ( plattnerite ) & tp@xmath16 & p@xmath17/mnm + + tio@xmath3 & anatase & ti@xmath13 & i@xmath18/amd + zro@xmath3 & baddeleyite & mp@xmath13 & p@xmath19/c@xmath20 + hfo@xmath3 & monoclinic(baddeleyite ) & mp@xmath13 & p@xmath19/c@xmath20 + .defect formation energy for group iv dopants at different concentrations . [ cols="<,^,^,^,^,^ " , ] the defect formation energy of doped ceria can also be calculated with regard to the dopant zo@xmath3 oxide ( note that some group iv dopants , such as pb , have multiple oxide stoichiometries ) . we chose to use zo@xmath3 type oxides since this corresponds to the stoichiometry of itself . furthermore , since elements like si have a large number of ( meta-)stable oxide polymorphs with the zo@xmath3 stoichiometry , and since temperature and pressure effects are not included in the _ ab initio _ calculations , we opt for a low pressure / temperature structure as found in reference . the oxide reference structures are shown in table [ table : suppl1_geombulkmetal ] . the defect formation energy , with regard to the oxides presented in table [ table : suppl1_geombulkmetal ] , is calculated as : @xmath21 with @xmath22 the total energy of the doped system , @xmath23 the total energy of bulk and @xmath24 the bulk energy of the oxide zo@xmath3 ( with z = c , si , ge , sn , pb , ti , zr , and hf ) . the resulting energies are shown in table [ table : metal4subst_energiesbis ] . the same general trends are observed as for the defect formation energy presented in the manuscript : a ) little concentration dependence , b ) large range of energies depending on the dopant element , c ) none of the dopants improve the stability of . the obtained energies , however , are lower due to the increased stability of the formed oxides . this means that for experiments , starting from the oxides , it will be easier to form group iv doped , then if one would start with the bulk dopant material . this research was financially supported by fwo - vlaanderen , project n@xmath25 g. @xmath26n . we also acknowledge the research board of ghent university . all calculations were carried out using the stevin supercomputer infrastructure at ghent university . y. she , q. zheng , l. li , y. zhan , c. chen , y. zheng , and x. lin , `` rare earth oxide modified cuo / ceo@xmath3 catalysts for the watergas shift reaction , '' _ int . j. hydrogen energy _ , 34[21 ] 89298936 ( 2009 ) . c. de leitenburg , a. trovarelli , f. zamar , s. maschio , g. dolcetti , and j. llorca , `` a novel and simple route to catalysts with a high oxygen storage capacity - the direct room - temperature synthesis of ceo@xmath3-zro@xmath3 solid solutions , '' _ j. chem . _ , 21 21812182 ( 1995 ) . m. manzoli , g. avgouropoulos , t. tabakova , j. papavasiliou , t. ioannides , and f. boccuzzi , `` preferential co oxidation in h@xmath3-rich gas mixtures over au / doped ceria catalysts , '' _ catal . today _ , 138[3 - 4 ] 239243 ( 2008 ) . d. e. p. vanpoucke , p. bultinck , s. cottenier , v. van speybroeck , and i. van driessche , `` density functional theory study of la@xmath31ce@xmath31o@xmath32 : disordered fluorite versus pyrochlore structure , '' _ phys . rev . b _ , 84[5 ] 054110 9pp . ( 2011 ) . m. paranthaman , a. goyal , f. list , e. specht , d. lee , p. martin , q. he , d. christen , d. norton , j. budai , and d. kroeger , `` growth of biaxially textured buffer layers on rolled - ni substrates by electron beam evaporation , '' _ physica c _ , 275[3 - 4 ] 266272 ( 1997 ) . s. oh , j. yoo , k. lee , j. kim , and d. youm , `` comparative study on the crack formations in the ceo@xmath3 buffer layers for ybco films on textured ni tapes and pt tapes , '' _ physica c _ , 308[1 - 2 ] 9198 ( 1998 ) . g. penneman , i. van driessche , e. bruneel , and s. hoste , `` deposition of ceo@xmath3 buffer layers and yba@xmath3cu@xmath33o@xmath34 superconducting layers using an aqueous sol - gel method,''pp . 501504 in key engineering materials , vol . 264268 , _ euro ceramics viii , pts 1 - 3 . _ edited by mandal , h and ovecoglu , l. turkish ceram soc ; european ceram soc , 2004 . y. takahashi , y. aoki , t. hasegawa , t. maeda , t. honjo , y. yamada , and y. shiohara , `` preparation of ybco coated conductor on metallic tapes using an mod process , '' _ physica c _ , 412 - 414[2 ] 905909 ( 2004 ) . k. knoth , b. schlobach , r. hhne , l. schultz , and b. holzapfel , `` la@xmath3zr@xmath3o@xmath35 and ce-gd-o buffer layers for ybco coated conductors using chemical solution deposition , '' _ physica c _ , 426 - 431[2 ] 979984 ( 2005 ) . n. van de velde , d. van de vyver , o. brunkahl , s. hoste , e. bruneel , and i. van driessche , `` ceo@xmath3 buffer layers for htsc by an aqueous sol - gel method - chemistry and microstructure , '' _ eur . j. inor . chem._,2010[2 ] 233241 ( 2010 ) . v. narayanan , p. lommens , k. de buysser , d. e. p. vanpoucke , r. huehne , l. molina , g. van tendeloo , p. van der voort , and i. van driessche , `` aqueous csd approach for the growth of novel , lattice - tuned la@xmath36ce@xmath37o@xmath38 epitaxial layers , '' _ j. mater . _ , 22[17 ] 84768483 ( 2012 ) . j. r. mcbride , k. c. hass , b. d. poindexter , and w. h. weber , `` raman and x - ray studies of ce@xmath40re@xmath36o@xmath41 , where re = la , pr , nd , eu , gd , and tb , '' _ j. appl . phys._,76[4 ] 24352441 ( 1994 ) . p. bera , k. r. priolkar , p. r. sarode , m. s. hegde , s. emura , r. kumashiro , and n. p. lalla , `` structural investigation of combustion synthesized cu / ceo@xmath3 catalysts by exafs and other physical techniques : formation of a ce@xmath37cu@xmath36o@xmath42 solid solution , '' _ chem . _ , 14[8 ] 35913601 ( 2002 ) . h. yamamura , h. nishino , k. kakinuma , and k. nomura , `` crystal phase and electrical conductivity in the pyrochlore - type composition systems , ln@xmath3ce@xmath3o@xmath35 ( ln = la , nd , sm , eu , gd , y and yb ) , '' _ j. ceram jpn _ , 111[12 ] 902906 ( 2003 ) . x. wang , j. a. rodriguez , j. c. hanson , d. gamarra , a. martnez - arias , and m. fernndez - garca , `` unusual physical and chemical properties of cu in ce@xmath37cu@xmath36o@xmath3 oxides , '' _ j. phys . chem . b _ , 109[42 ] 1959519603 ( 2005 ) . a. tiwari , v. m. bhosle , s. ramachandran , n. sudhakar , j. narayan , s. budak , and a. gupta , `` ferromagnetism in co doped ceo@xmath3 : observation of a giant magnetic moment with a high curie temperature , '' _ appl . _ , 88[14 ] 142511 3pp . ( 2006 ) . a. nakamura , n. masaki , h. otobe , y. hinatsu , j. wang , and m. takeda , `` defect - fluorite oxides m@xmath43ln@xmath44o@xmath45(ln = lanthanide ; m = hf , zr , ce , u , th ) : structure , property , and applications , '' _ pure appl . _ , 79[10 ] 16911729 ( 2007 ) . b. vodungbo , y. zheng , f. vidal , d. demaille , v. h. etgens , and d. h. mosca , `` room temperature ferromagnetism of co doped ceo@xmath47 diluted magnetic oxide : effect of oxygen and anisotropy , '' _ appl . _ , 90[6 ] 062510 3pp . ( 2007 ) . r. k. singhal , p. kumari , s. kumar , s. n. dolia , y. t. xing , m. alzamora , u. p. deshpande , t. shripathi , and e. saitovitch , `` room temperature ferromagnetism in pure and co- and fe - doped ceo@xmath3 dilute magnetic oxide : effect of oxygen vacancies and cation valence , '' _ _ , 44[16 ] 165002 6pp . ( 2011 ) . b. m. reddy , a. khan , p. lakshmanan , m. aouine , s. loridant , and j .- c . volta , `` structural characterization of nanosized ceo@xmath3-sio@xmath3 , ceo@xmath3-tio@xmath3 , and ceo@xmath3-zro@xmath3 catalysts by xrd , raman , and hrem techniques , '' _ b _ , 109[8 ] 33553363 ( 2005 ) . b. reddy and a. khan , `` nanosized ceo@xmath3sio@xmath3 , ceo@xmath3tio@xmath3 , and ceo@xmath3zro@xmath3 mixed oxides : influence of supporting oxide on thermal stability and oxygen storage properties of ceria , '' _ catal asia _ , 9[3 ] 155171 ( 2005 ) . e. rocchini , a. trovarelli , j. llorca , g. w. graham , w. h. weber , m. maciejewski , and a. baiker , `` relationships between structural / morphological modifications and oxygen storage redox behavior of silica - doped ceria , '' _ j. catal . _ , 194[2 ] 461478 ( 2000 ) . e. rocchini , m. vicario , j. llorca , c. de leitenburg , g. dolcetti , and a. trovarelli , `` reduction and oxygen storage behavior of noble metals supported on silica - doped ceria , '' _ j. catal . _ , 211[2 ] 407421 ( 2002 ) . r. lin , m .- f . zhong , z .- yan , g .- y . liu , and w .- p . liu , `` comparative study of cuo / ce@xmath49sn@xmath50o@xmath3 , cuo / ceo@xmath3 and cuo / sno@xmath3 catalysts for low - temperature co oxidation , '' _ appl . , a _ , 255[2 ] 331336 ( 2003 ) . d. a. andersson , s. i. simak , n. v. skorodumova , i. a. abrikosov , and b. johansson , `` redox properties of ceo@xmath3mo@xmath3 ( m = ti , zr , hf , or th ) solid solutions from first principles calculations , '' _ appl . _ , 90[3 ] 031909 3pp . ( 2007 ) . y. tang , h. zhang , l. cui , c. ouyang , s. shi , w. tang , h. li , j .- s . lee , and l. chen , `` first - principles investigation on redox properties of @xmath51-doped @xmath31 @xmath52 , '' _ phys . b _ , 82[12 ] 125104 9pp . ( 2010 ) . m. yashima , `` crystal structures of the tetragonal ceria zirconia solid solutions ce@xmath36zr@xmath37o@xmath3 through first principles calculations ( @xmath53 ) , '' _ c _ , 113[29 ] 1265812662 ( 2009 ) . d. e. p. vanpoucke , s. cottenier , v. van speybroeck , p. bultinck , and i. van driessche , `` tuning of ceo@xmath3 buffer layers for coated superconductors through doping , '' _ appl . _ , 260 3235 ( 2012 ) . d. a. andersson , s. i. simak , b. johansson , i. a. abrikosov , and n. v. skorodumova , `` modeling of ceo@xmath3 , ce@xmath3o@xmath33 , and ceo@xmath54 in the lda+u formalism , '' _ phys . b _ , 75[3 ] 035109 6pp . ( 2007 ) . m. v. ganduglia - pirovano , a. hofmann , and j. sauer , `` oxygen vacancies in transition metal and rare earth oxides : current state of understanding and remaining challenges , '' _ surf . sci . reports _ , 62[6 ] 219270 ( 2007 ) . y. tsuru , y. shinzato , y. saito , m. shimazu , m. shiono , and m. morinaga , `` estimation of linear thermal expansion coefficient from cohesive energy obtained by ab - initio calculation of metals and ceramics , '' _ j. ceram _ , 118[1375 ] 241245 ( 2010 ) . s. l. dudarev , g. a. botton , s. y. savrasov , c. j. humphreys , and a. p. sutton , `` electron - energy - loss spectra and the structural stability of nickel oxide : an lsda+u study , '' _ phys . b _ , 57[3 ] 15051509 ( 1998 ) . c. loschen , j. carrasco , k. m. neyman , and f. illas , `` first - principles lda@xmath55u and gga@xmath55u study of cerium oxides : dependence on the effective u parameter , '' _ phys . b _ , 75[3 ] 035115 8pp . ( 2007 ) . m. a. blanco , a. m. pendas , e. francisco , j. m. recio , and r. franco , `` thermodynamical properties of solids from microscopic theory : applications to mgf@xmath3 and al@xmath3o@xmath33 , '' _ theochem - j . _ , 368 245255 ( 1996 ) .	fluorite doped with group iv elements is studied within the dft and dft+u framework . concentration dependent formation energies are calculated for ( z= c , si , ge , sn , pb , ti , zr , hf ) with @xmath0 and a roughly decreasing trend with ionic radius is observed . the influence of the valence and near valence electronic configuration is discussed , indicating the importance of filled @xmath1 and @xmath2 shells near the fermi level for all properties investigated . a clearly different behavior of group iva and ivb dopants is observed : the former are more suitable for surface modifications , the latter are more suitable for bulk modifications . + for the entire set of group iv dopants , there exists an inverse relation between the change , due to doping , of the bulk modulus and the thermal expansion coefficients . hirshfeld - i atomic charges show that charge transfer effects due to doping are limited to the nearest neighbor oxygen atoms .
using organic molecules as functional units for electronic apparatus application is an interesting goal of nanoelectronic devices.@xcite the common and important function of these devices is that the current can be controlled effectively . in the last several years , many experimental and theoretical works were carried out to study the transport properties through a single molecule , or even to design the molecular electronic devices . @xcite at present , people have realized two major approaches to control molecular transport . one is through the conformational change in the molecule , the other is through the external transverse field to switch the molecule from an on to an off state . many authors have focused their attention on the former one for a long time.@xcite despite the conformational change in the molecule can be achieved by using the electric or light fields , its operational frequency is low . now , the attention is transferred to the latter one due to its high operation frequency . several experiments have identified their feasibility.@xcite currently , the @xmath0-conjugated organic oligomers and polymers are the subject of considerable research interest in the organic semiconductors . the organic semiconductors can be employed as active layers in the field effect transistors ( fet).@xcite the gate - bias controlled molecular transistor is successfully achieved experimentally @xcite from perylene tetracarboxylic diimide ( ptcdi ) , a redox molecule . recently , the current behavior of the single molecule has received increasing attention . single organic oligomers such as pentathienoacene ( pta ) , pentacene , perylene and so on are all the key objects in the theoretical study . but at present , few theoretical work sheds light on the organic molecular transistor controlled by the transverse field . the rigorous treatment of the molecular device in theory calls for the combination of the theory of quantum transport with the first - principles calculations of the electronic structure in the self - consistent scheme . in this paper , we use the density functional theory ( dft ) and non - equilibrium green s function to study the transverse field effect ( tfe ) on current transport of the single organic oligomer . the retarded green s function of the molecule is expressed as follows @xmath1 where @xmath2 and @xmath3 are the overlap matrix and fock matrix of the molecule part , respectively . @xmath4 ( @xmath5 ) is the retarded self - energy of the left ( right ) electrode . it should be emphasized that the fock matrix is obtained after the density matrix is obtained . the density matrix of the open system is the essential function of the whole self - consistent scheme . it can be achieved by the keldysh green s function @xmath6,\ ] ] @xmath7 with the advanced green s function @xmath8 , the broadening function of the left ( right ) lead @xmath9 ( @xmath10 ) . the fermi distribution function of the left ( right ) lead @xmath11 ( @xmath12 ) is expressed @xmath13 with @xmath14 , @xmath15 . @xmath16 is fermi level of the bulk au . in our work , @xmath16 is -5.1 ev which is an adjusted parameter around its work function ( 5.31 ev ) for explaining experimental results.@xcite if the transverse field is applied perpendicular to the transport direction , the corresponding potential energy term is included in the fock operator @xmath17 our model is illustrated in fig . 1 with the molecule attached by the gold electrodes from the both sides and the transverse electric field confined inside the molecule region . the potential zero point is set at the coordinate origin , the center of the line connecting two sulfur atoms and the whole molecule is not symmetric to the au - s bond . since the transverse field is localized in the molecule , its edge effect is very weak so that it almost has no effect on the electrodes . in the calculation , the applied longitudinal voltage points to the current direction shown in fig . the molecule is chemisorbed on the gold contacts by sulfur atoms . the sulfur atom sits on the hollow position of three nearest - neighbor surface gold atoms . the perpendicular distance between the sulfur atom and the gold fcc ( 111 ) surface plane is 2.0 , an usually acceptable distance . the temperature effect is not distinct for the short molecule , so we assume zero temperature in our calculation for simplicity . ] the density matrix for the open system should be obtained self - consistently . in order to achieve this goal , we extended the inner loop in the standard quantum chemistry software gaussian03 @xcite to the loop composed of the lead - molecule - lead open system under bias . at the beginning , the self - consistent procedure starts from a guess for the density matrix of the open system , which may be obtained from the converged density matrix of gaussian03 calculation for the isolated molecule . we feedback the density matrix to the gaussian s main program as a subroutine to obtain the new density matrix . the iterations continue until the density matrix converges to the acceptable accuracy ( usually less than @xmath18 ) . then the density matrix is used to evaluate the transmission function ( t ) , the terminal current , the electron number and the density of states ( dos ) of the open system . @xcite in the calculations , we adopt dft with b3pw91 exchange - correlation potential and lanl2dz basis to evaluate the electronic structure and the fock matrix . the basis set associates with the effective core potential ( ecp ) , which are specially suited for the fifth - row ( cs - au ) elements including the darwin relativistic effect . the sulfur atom in thiophene is sp@xmath19 hybridized , and its p - orbital provides two electrons to the @xmath0 system . the linearly condensed thiophene molecules possess the extended @xmath0 conjugation and the high planarity . of the cyclic single organic molecules , the thiophene - based compound is one of the promising class of organic materials . pta is attractive , due to the stability of the thiophene ring and the good planarity.@xcite fig . 2 gives the dos and t of pta with the different transverse gate biases . the position of the broadened levels in equilibrium for the open system is determined by the singular points of the green s function , obtained from equation @xmath20 . the molecular levels are modulated by the gate bias applied to the molecule with homo ( lumo ) position at @xmath21 ( @xmath22 ) ev for @xmath23 v ( a ) ; @xmath24 ( @xmath25 ) ev for @xmath26 v ( b ) ; @xmath27 ( @xmath28 ) ev for @xmath29 v ( c ) ; @xmath30 ( @xmath31 ) ev for @xmath32 v ( d ) , and @xmath33 ( @xmath34 ) ev for @xmath35 v ( e ) . with decrease of the positive gate bias , the separation between homo and fermi level increases , and the separation between lumo and fermi level decreases . for the positive gate bias @xmath36 , the fermi level is close to homo , and pta is the p - type or hole conduction molecule , while for the big negative bias pta is the n - type or electron conduction molecule . for large positive bias @xmath35 v , the homo , close to @xmath16 , contributes to the initial rise of the current under the small applied voltage ( 1.0 v or so ) in the longitudinal direction . meanwhile lumo is close to @xmath16 in the case of large negative bias @xmath23 v , it is responsible for the molecular conduction ( fig . 3 ) . fig . 2 illustrates that in fig . 3 the gate bias successfully modulates the i - v characteristics of the pta . at the voltage @xmath37 v , due to the contribution from homo , the positive gate bias achieves the bigger molecular current and at the voltage @xmath38 v , as lumo enters the voltage window and contributes the molecular conduction , the inverse order of current appears . the inset illustrates the @xmath39 electron number deviation from the equilibrium state as a function of bias . at applied voltage @xmath37 v , for gate bias @xmath23 v , the electron is responsible for the conduction ( @xmath40 ) and electron number rises slowly , while for @xmath41 v , the hole is responsible for the conduction ( @xmath42 ) , so electron number descends ( the electron flows out of the molecule to the lead ) . at applied voltage @xmath38 v , the molecule enters the electron- and hole- hybrid conduction region , companying the rapid rise of electron number , which means the electron contribution dominates the conduction . v ( a ) , @xmath43 v ( b ) , 0 v ( c ) , + 1.95 v ( d ) , and + 3.90 v ( e ) . the vertical line denotes the position of fermi level . [ fig2 ] ] electron number deviation from the equilibrium state with the modulation of bias . [ fig4 ] ] similar to pta , the good tfe is achieved by the other fused - ring thiophene molecules . 4(a ) and ( b ) show the i - v curves of the 3-fused - ring thiophene and 7-fused - ring thiophene controlled by the gate bias . for example , at applied voltage @xmath44 v , for the 3-fused - ring thiophene molecules , @xmath45 , @xmath46 ; for the 7-fused - ring thiophene , @xmath47 , @xmath48 . meanwhile at the same bias for pta , @xmath49 , @xmath50 . the longer fused - ring thiophene molecules have better tfe property . although the early experiment @xcite reported only weak gate effect on the molecular current observed for the single benzene - dithiol molecule . our calculation found that the tfe can be improved by the longer polyacene - dithiol molecules , which present a good gate - bias controlled molecular transistor property . dos and t of the single pdt , naphthalene - dithiol , and anthracene - dithiol in equilibrium without the longitudinal and transverse field are shown in fig . 5 , where both homo and lumo are close to @xmath16 with increase of the aromatic phenyl ring . the homo ( lumo ) is @xmath51 ( @xmath52 ) ev for the single pdt , @xmath53 ( @xmath54 ) ev for the naphthalene - dithiol , and @xmath55 ( @xmath56 ) ev for the anthracene - dithiol . the fact predicts that if the number of aromatic phenyl ring is increased to more than 3 , homo and lumo will be closer to @xmath16 , and the tfe will be more apparent . for the pentacene , @xmath57 2 ev and @xmath16 is almost in the middle of h - l gap , it has apparent tfe in the small bias . the anthracene - dithiol is the n - type conduction molecule in the small voltage region , its lumo is closer to @xmath16 than homo [ see fig . its molecular levels can be modulated obviously by the gate - bias . with the gate - bias increased by increment @xmath58 v from @xmath59 v to @xmath60 v , the homo of anthracene - dithiol leaves @xmath61 by values @xmath62 ev , @xmath63 ev , @xmath55 ev , @xmath64 ev , @xmath65 ev ; and lumo approaches to @xmath61 by values @xmath66 ev , @xmath67 ev , @xmath56 ev , @xmath68 ev , and @xmath69 ev , respectively . with the gate bias @xmath70 v , the separation between @xmath16 and lumo , @xmath71 ev , which leads the large current at small bias . the i - v curves of anthracene - dithiol controlled by the gate bias are shown in fig . the current magnitude of the anthracene - dithiol is one order smaller than the one of the fused - ring thiophene molecules , which is in favor of the low - power device . the molecular transistor illustrates the effective current separation under control of the positive gate bias . for the applied voltage @xmath72 v , @xmath73 , @xmath74 ; while for @xmath75 v applied voltage , @xmath76 , @xmath77 . we notice that in the voltage range from @xmath72 v to @xmath75 v , the current under the transverse field of @xmath59 v is bigger than that at @xmath29 v , which is not found for fused - ring oligothiophene molecules . the gate - controlled current separation for the negative bias is not effective , we have to use its positive - bias controlled function . for the anthracene - dithiol , the transverse field of @xmath59 v makes homo close to @xmath78 ev , accompanying the sudden rise of current around @xmath75 v voltage , so its current is larger than the one for @xmath29 where neither homo nor lumo enters the voltage window . the inset of fig . 6 illustrates the electron number deviation from equilibrium . for the open system without the transverse bias , the @xmath39 electron number has almost no variation below the applied voltage @xmath79 v , since both homo and lumo keep away from @xmath16 in the voltage range . with increase of gate bias , lumo is shifted to @xmath16 gradually , and the electrons entering the molecule are more than that leaving the molecule , which makes the electron number increase . with decrease of bias , homo is close to @xmath16 gradually , which makes the electron number decrease . ] with increase in the number of aromatic phenyl ring , the tfe is improved . we found tfe of the naphthalene - dithiol is better than that of the single pdt , but not better than that of the anthracene - dithiol . for the single pdt , under bias @xmath72 v , @xmath80 , @xmath81 , while under bias -2.0 v , @xmath82 , @xmath83 . the control of the molecular current by the gate bias is not discernible , which is consistent with the experiment.@xcite the i - v curves of the pentacene under the different gate bias are shown in fig . 7 . under bias @xmath72 v , the molecular device presents the obvious current modulation by the gate - bias : @xmath84 ; @xmath85 ; @xmath86 . the molecule illustrates the five times of the current modulation rate of the anthracene - dithiol under the same gate bias . considering the stability of the structure , maybe the anthracene - dithiol is still a better candidate for the molecular transistor . the above tfe comes from the fact that with the increase in transverse field for the polyacene - dithiol , it will make energy levels lower and for the fused ring thiophene , with the increase in the transverse field , it will make energy levels higher . it should be noted that the energy level shift is determined by the choose of the potential zero point , which is set at the center of the line connected by two sulfur atoms in our model . if the potential zero point deviates from this point , the expression of chemical potential of left and right leads should be rectified : @xmath87 , and @xmath88 . @xmath89 is the electrical potential difference between the old zero point and the new one . however , the choose of zero point has no effect on the calculation results . we use the self - consistent method based on the dft and non - equilibrium green s function to simulate molecular transport . in terms of gaussian03 , the electronic structures of the molecular device and the macroscopic leads are calculated on an equal footing . at the same time , the self - consistent iteration cycle is extended from the local molecule to the open system , by inserting the calculation of the density matrix of the open system as a subroutine . our self - consistent results show that the long - length organic molecule can achieve better transport characteristics . our investigation proves that both the polyacene - dithiol molecules and fused - ring oligothiophene molecules can be made as the high - frequency molecular transistors controlled by the transverse field . the theoretical results suggest the related experiments about the molecular devices . this work is supported by the national natural science foundation of china ( nsfc ) under projects 90206031 and 10574024 . the authors gratefully acknowledge sr8000 supercomputer resources from the center for computational materials science of the institute for materials research , tohoku university . 99 a. nitzan and m.a . ratner , science * 300 * , 1384 ( 2003 ) . reed and t. lee , eds . , _ molecular nanoelectronics _ , ( american scientific publishers , ca , 2003 ) . reed , c. zhou , c.j . muller , t.p . burgin , and j.m . tour , science * 278 * , 252 ( 1997 ) . gittins , d. bethell , d.j . schriffrin , and r.j . nichols , nature * 408 * , 67 ( 2000 ) . cui , a. primak , x. zarate , j. tomfohr , o.f . sankey , a.l . moore , t.a . moore , d. gust , g. harris , s.m . lindsay , science * 294 * , 571 ( 2001 ) . b. larade , j. taylor , q.r . zheng , h. mehrez , p. pomorski , and h. guo , phys . b * 64 * , 195402 ( 2001 ) . zhitenev , h. meng , and z. bao , phys . * 88 * , 226801 ( 2002 ) . j. park , a.n . pasupathy , j.i . goldsmith , c. chang , y. yaish , j.r . petta , m. rinkoski , j.p . sethna , h.d . abruna , p.l . mceuen , and d.c . ralph , nature * 417 * , 722 ( 2002 ) . w. liang , m.p . shores , m. bockrath , j.r . long , and h. park , nature * 417 * , 725 ( 2002 ) . smit , y. noat , c. untiedt , n.d . lang , m.c . van hemert , and j.m . van ruitenbeek , nature * 419 * , 906 ( 2002 ) . j. reichert , r. ochs , d. beckmann , h.b . weber , m. mayor , and h.v . l@xmath90hneysen , phys . rev . lett . * 88 * , 176804 ( 2002 ) . smit , c. untiedt , g. rubio - bollinger , r.c . segers , and j.m . van ruitenbeek , phys . * 91 * , 076805 ( 2002 ) . d. dulic , s.j . van der molen , t. kudernac , h.t . jonkman , j.j.d . de jong , t.n . bowden , j. van esch , b.l . feringa , and b.j . van wees , phys . * 91 * , 207402 ( 2003 ) . b. xu and n.j . tao , science * 301 * , 1221 ( 2003 ) . b. xu , x. xiao and n.j . tao , j. am . soc . * 125 * , 16164 ( 2003 ) . x. xiao , b. xu , and n.j . tao , nano lett . * 4 * , 267 ( 2003 ) . heller , iee proc . * 130 * , 209 ( 2000 ) . liu , k. hashimoto and a. fujishima , nature * 347 * , 658 ( 1990 ) . kornilovitch , a.m. bratkovsky , and r.s . williams , phys . b * 66 * , 245413 ( 2002 ) . reed , j.m . tour , sci . am . * 282 * , 86 ( 2000 ) . seminario , p.a . derosa , and j.l . bastos , j. am . soc . * 124 * , 10266 ( 2002 ) . r. pati and s.p . karna , phys . b * 69 * , 155419 ( 2004 ) . a. ghosh , t. rakshit and s. datta , nano lett . * 4 * , 565 ( 2004 ) . hou , j.x . zhang , r. li , j. ning , r.s . han , z.y . shen , x.y . zhao , z.q . xue and q.d . wu , nanotechnology * 16 * , 239 ( 2005 ) . m. kondo , t. tada , and k. yoshizawa , chem . * 412 * , 55 ( 2005 ) . k. xiao , y. liu , t. qi , w. zhang , f. wang , j. gao , w. qiu , y. ma , g. cui , s. chen , x. zhan , g. yu , j. qin , w. hu , and d. zhu , j. am . soc . * 127 * , 13281 ( 2005 ) . y. sun , y. liu and d. zhu , * 15 * , 53 ( 2005 ) b. xu , x. xiao , x. yang , l. zang and n. tao , j. am . soc . * 127 * , 2386 ( 2005 ) . damle , a.w . ghosh , s. datta , chem . phys . * 281 * , 171 ( 2002 ) . f. jiang , y.x . zhou , h. chen , r. note , h. mizuseki , and y. kawazoe , phys . b * 72 * , 155408 ( 2005 ) . frisch , et al . , gaussian03 , revision b. 04 , ( gaussian , inc . , pittsburgh pa , 2003 ) . j .- o lee , g. lientschnig , f. wiertz , m. struijk , r.a.j . janssen , r. egberink , d.n . reinhoudt , p. hadley , and c. dekker , nano lett . * 3 * , 113 ( 2003 ) .	we use a self - consistent method to study the current of the single molecular transistor modulated by the transverse gate - bias in the level of the first - principles calculations . the numerical results show that both the polyacene - dithiol molecules and the fused - ring oligothiophene molecules are the potential high - frequency molecular transistor controlled by the transverse field . the long molecules of the polyacene - dithiol or the fused - ring thiophene are in favor of realizing the gate - bias controlled molecular transistor . the theoretical results suggest the related experiments .
the ability to control ultra - cold atom - atom interactions has opened the door to a wide variety of fundamental and applied studies , including the production of ultra - cold molecules @xcite , simulations of condensed matter phenomena @xcite , and quantum information processing @xcite . the tools that have been central to this development include designer atom traps , for example , optical lattices @xcite , and controllable scattering resonances such as a magnetic feshbach resonance @xcite . both of these can be used to manipulate the two - body scattering process , thus affecting the strength of the interaction , the nature of the resulting two - body states , and more general many - body phenomena . examples include confinement - induced resonances @xcite , bound - states with repulsive interactions @xcite , and feshbach resonances in band structures @xcite . a particular example that we have explored previously is a trap - induced resonance ( tir ) that occurs as a result of interaction between atoms that are confined to spatially separated harmonic traps @xcite . this happens as a consequence of a molecular bound state that becomes resonant with the vibrational state of the separated atoms due to a quadratic rise in the light shift when the two atoms approach one another , as shown schematically in fig . [ tirschematic ] . a strong resonance can occur when the confinement of the wave packet in the trap is on the order of ( or smaller than ) the free - space scattering length . this @xmath2-wave resonance is analogous to a higher - partial wave shape resonance occurring in free space , but here the tunneling barrier arises from the trap rather than from an angular momentum centrifugal barrier . because of this tunneling , the interaction occurs at negative energy " values with respect to the free - particle @xmath2-wave scattering threshold . in previous work on the tir in cesium @xcite , whose large scattering length gives rise to a strong resonance , stock _ et al . _ extracted the scattering length at negative energy for the single channel case by an explicit integration of the radial schrdinger equation at negative energies using the numerov method . such a procedure has limited utility ; the solutions are unstable since the wave function blows up in the tunneling barrier . the situation gets very complicated as soon as there is more than one channel . a proper numerical technique must ensure that open channels are propagated along with the exponentially decaying closed channels while maintaining accuracy to relevant digits . most coupled channel codes that incorporate such situations ( i.e. propagating closed and open channels ) , eventually drop the closed channels beyond certain radius since they are only interested in open channels . on the other hand , the problem of tir considered in this paper requires us to integrate to large enough radius for determining a good asymptotic log derivative ( for both open and closed channels ) . we emphasize that we are not allowed to drop any channels , since in the end , we are required to extract the asymptotic log derivatives for both open as well as closed channels . to remedy this , we consider here _ analytic models _ of the multichannel @xmath0-matrix . simplified analytic models have been employed in previous studies of ultra - cold collisions and scattering resonances . julienne and gao have predicted the properties of feshbach resonances based on the analytic properties of the van der waals long range potential @xcite . have used analytic models to describe the interplay of open and closed channels in the context of feshbach resonances associated with a large background scattering length @xcite . these analytic models , while simple in nature , are able to encapsulate the necessary physics in just a few parameters . these parameters can then be incorporated into building model many - body hamiltonians , an example being the two - channel model used for describing resonance superfluidity in a two component fermi gas @xcite . our goal in this article is to develop an analytic model that can be used to study the tir for cs atoms trapped in an optical lattice . in sec . ii we review the basic physics that gives rise to the tir , its application in cs , and show the limitations of direct numerical solutions , even for single channel scattering . section iii contains the heart of our new results . we review the basic resonant scattering phenomena and how they are modeled analytically in the @xmath0-matrix . we then apply this to determine expressions for the negative - energy scattering length in a nontrivial multichannel scattering process , relevant to an experimental observation of the tir . we summarize our results in sec . hyperfine manifold of @xmath1cs ( not to scale ) . the solid ( dashed ) lines correspond to the levels in the absence(presence ) of the an external magnetic field . the magnetic quantum numbers and the corresponding level - labellings are shown . ] we consider the scattering of two @xmath1cs atoms in their 6s@xmath3 electronic ground state , trapped in an optical lattice . the zeeman hyperfine structure of this manifold is shown in the fig . [ levels ] with magnetic sub - levels labeled as @xmath4 for convenience . henceforth , all two - atom scattering channels will be denoted by the relevant pair of these sub - levels . to begin with , we consider the scattering in the @xmath5 channel . this is motivated by studies of controlled collisions via spin - dependent transport in polarization gradient lattices @xcite . in these spin states , two atoms that are separated by @xmath6 in a lin - perp - lin optical lattice can be transported into the same well in a lin - parallel - lin optical lattice via a rotation of the laser polarization . by angular momentum conservation , because these are stretched states " , and ignoring small spin - dipolar and second - order spin - orbit coupling @xcite , the dominant @xmath2-wave collision does not couple this channel to any other channel . the result is an elastic phase shift that can be used to implement an entangling two - qubit logic gate . in addition , as the wells approach one another , there will be a tir that can strongly affect the two - atom interaction @xcite . channel in units of bohr radius @xmath7 as a function of scattering energy @xmath8 . ] the properties of the tir follows from a simple model of the two - atom system . we express the hamiltonian for the atoms in the @xmath5 scattering channel in center - of - mass coordinates , @xmath9 , and relative coordinates , @xmath10 , as @xmath11 where @xmath12 is the reduced mass , @xmath13 is the separation of the traps , and @xmath14 is the inter - atomic potential . in principle , the tir can be seen by diagonalizing the above schrdinger equation using the precise cs@xmath15 molecular potential projected on @xmath5 for @xmath14 . this is a non - trivial task , however , since there is a huge separation of length scales between the molecular potential and the external trapping potential , and the displacement of the trap from the zero of the relative - coordinate makes the system anisotropic . instead , we treat the molecular potential through a contact pseudo - potential @xcite , @xmath16 here @xmath17 is the energy - dependent @xmath2-wave scattering length , determined by direct numerical integration of the schrdinger equation based on the @xmath2-wave scattering phase shift of the known cs@xmath15 molecular potential in the _ absence _ of the trap , according to @xmath18 . the energy is then chosen self - consistently to solve the schrdinger equation , including both the boundary conditions at short - range due to the atomic interaction and at long - range due to the trap @xcite . for sufficient separation between the traps , the lowest energy eigenstates drops below the threshold of the molecular potential . thus the negative energy " scattering states that are inaccessible in free space , become opened by the trapping potential . the scattering length at positive energies for the @xmath5 channel , calculated using a numerical solution to the radial schrdinger equation based on the well - established @xmath1cs dimer potential , is shown in fig . [ rvalues ] . a resonance exists at @xmath19 @xmath12k ( here and throughout , energy is measured in temperature units ) due to a bound state very close to zero energy . finding the scattering length at negative energies via equivalent numerical integration is highly unstable , as the wave function blows up in the classically forbidden region . even if one manages to do it , it is necessary that the integration be sufficiently stable for large @xmath20 , well into the asymptotic region of the potential , in order to extract a meaningful scattering length @xcite . in addition , because we are in the neighborhood of a scattering resonance near zero energy , the strong variation of the scattering length with energy makes the requirement for a robust numerical solution even more demanding . to address these problems , we develop analytic models that will allow us to calculate the scattering matrix below threshold . in doing so , we will extend the method to a more complicated process that occurs for the multichannel scattering case . near - threshold scattering is dominated by resonant phenomena . such resonances can arise from a variety of different physical mechanisms , and thereby affect the form of the analytic model . we identify the nature of the resonance based on our understanding of the physical processes and the location of the poles in the @xmath0-matrix . away from resonance , the scattering properties are smooth functions and therefore can be modeled by a few free parameters . the total @xmath0-matrix thus factors into resonant and background contributions , @xmath21 for @xmath2-wave scattering on the @xmath5 channel . the circles represent numerical data from coupled channel calculation , the green curve is the analytical fit using eq . ( [ linearphase ] ) , and the red curve is the analytical fit using eq . ( [ effectiverange ] ) . ] to negative energies.the circles represent data from numerical integration of the schrdinger equation . ] for the scattering on the single channel @xmath5 , the resonant behaviour can arise only from a bound state or a virtual - bound state near threshold ; any other type of resonance such as a feshbach will require the inclusion of other channels . a direct numerical integration gives a bound state with a binding energy @xmath22 @xmath23 . thus , the resonant part of the single channel @xmath0-matrix can be written as @xcite , @xmath24 where @xmath25 , representing a pole of the @xmath0-matrix in the @xmath26-plane on the positive imaginary axis . there is no need to single out the other bound states , as they are energetically too far away , and their effect is absorbed in the background part . this remaining part can be written in the low energy limit @xmath27 as , @xmath28 where @xmath29 is the background scattering length that encapsulates the effect of all other non - resonant processes , including other deeply bound states . now we can write the complete @xmath0-matrix element analytically with just one free parameter , @xmath29 , the value of which can be determined by fitting one positive low energy point to the equation @xmath30 in fig . [ phasefit ] we plot the scattering phase shift as a function of the scattering energy , obtained via the full coupled channels calculation and compare it with the one obtained from eq . ( [ linearphase ] ) . we see good agreement at low energy but for energies beyond 10 @xmath12k a slight difference is noticed . this is expected since the linear form of the background phase shift is only valid at low energies . to remedy this problem , we use a higher order expansion for the background part given by @xmath31 where we have added an additional free parameter , @xmath32 , the effective range . as before , we determine both the parameters by fitting two low energy data points . the inset of fig . [ phasefit ] shows excellent agreement of this improved model with the numerical solution . given the form of the s - matrix , we can predict the scattering properties in the @xmath5 channel at negative energy values by performing an analytic continuation of the @xmath0-matrix to the imaginary @xmath26-axis . from this one can consistently define the scattering length at negative energies @xmath33 by @xmath34 in fig . [ scatfit ] we plot the scattering lengths obtained from the analytic procedure discussed above . these agree with the direct numerical integration just below threshold @xcite . we now turn to a more complex situation : collisions between @xmath35 and @xmath36 or the @xmath37 channel . this is motivated by the following experimental considerations . controlled collisions via spin - dependent transport in polarization gradient optical lattices @xcite is hampered by inhomogeneous broadening arising from unwanted real or fictitious magnetic fields ( due to elliptically polarized light at the atomic position ) @xcite . this is a particularly deleterious effect for the @xmath5 states that see a strongly varying difference in their optical potentials along the transport . in a lin - angle - lin optical lattice at very large detunings , atoms in the @xmath38 and @xmath39 states experience almost the same shift due to the fictitious magnetic field . any residual broadening is due to the finite detuning effects ( giving a differential scalar light shift ) and the true magnetic field inhomogeneity . controlled collisions in the @xmath37 channel thus offer the advantage of higher degrees of coherence , with potential applications to quantum information processing . to treat scattering with the incoming @xmath37 channel , we must account for the exchange interaction , which leads to spin - changing collisions that preserve the total projection of angular momentum along a quantization axis . in this case , the @xmath2-wave collisions couple the @xmath37 channel to the @xmath40 , @xmath41 , @xmath42 , and @xmath43 channels . at low energies , small compared to the hyperfine splitting , @xmath42 and @xmath43 are energetically closed . moreover , in the presence of any positive magnetic field @xmath44 , the channel @xmath40 shifts to a higher energy compared to the @xmath37 channel , as depicted in fig . [ potschematic ] . for energies that are smaller than this shift , @xmath40 channel is also closed . the movement of this channel from below to above the threshold can lead to a feshbach resonance that strongly affects the scattering process , as discussed below . . at @xmath45 , the bound state corresponding to the @xmath40 channel crosses the @xmath37 channel threshold . the @xmath40 channel becomes closed for @xmath44 . ] the scattering phase shift for the @xmath37 channel is calculated by a full coupled - channels calculation . in fig . [ aophase ] we plot @xmath46 as a function of scattering energy @xmath8 and the external magnetic field @xmath47 in the range of a few hundred mg . as is seen in this figure , there is a scattering resonance along the dashed line where @xmath48 . also , since the resonance moves monotonically upwards in energy as a function of the @xmath47 field , it is clear that this resonance is a feshbach resonance . we confirm this by calculating the bound state energy and find that it changes sign at approximately @xmath49 mg as shown in fig . [ boundstate ] . there also exists a threshold @xmath50 corresponding to the opening of the @xmath40 channel ( shown by the white solid line in fig . [ aophase ] ) . across this threshold the number of open channels changes by one , as reflected in the abrupt change in the @xmath51 dependence of the scattering properties . this is indicated in fig . [ jumpinphase ] where we plot the scattering phase shift as a function of energy for various values of the @xmath47 field . channel as a function of energy and magnetic field , plotted as surface plot of @xmath46 . the solid line corresponds to the boundary in the ( e , b ) plane that separates the region where channel @xmath52 is closed and open . the dashed line corresponds to the points where @xmath48 . the color scheme used is linear with dark red corresponding to the value @xmath53 and dark blue to @xmath54 . ] channel bound state which moves up with b - field . inset shows the shift of the bound state due to dressing by the @xmath37 channel ] based on the above understanding of the mechanisms that lead to scattering resonances , our goal is to build an analytic model that will allow us to predict the scattering lengths at negative energies where the tir is predicted to occur . to begin with , we will again assume that the diagonal element of the @xmath0 matrix can be modeled as @xmath55 where @xmath56 is a smooth function describing the background contribution and @xmath57 is the contribution arising from the feshbach resonance . as before , the resonance leads to a pole in the @xmath0-matrix . in fact it can be shown that every element of the multichannel @xmath0-matrix has a pole corresponding to this resonance @xcite . also , since the location of the resonance moves upwards in energy almost linearly as a function of the @xmath47 field , it is fair to assume that the bound state solely resides on the @xmath40 channel and is shifted by an amount @xmath58 due to coupling to the @xmath37 channel . therefore without considering the off - diagonal elements of the @xmath0-matrix , the resonant part can be modeled by a pure breit - wigner pole , @xmath59 so that that feshbach contribution to the phase shift is @xmath60.\label{feshphase}\ ] ] here @xmath61 is the location of the bare bound state of the @xmath40 channel , @xmath62 is the width of the feshbach resonance and represents the location of the pole of @xmath63 along the imaginary axis in the complex @xmath8 plane . this model is restricted to energies below the @xmath40 threshold for a finite magnetic field , @xmath50 . the discontinuity in the scattering phase shift seen in fig . [ jumpinphase ] shows that a different model is required as additional scattering channels are opened . we leave that problem for future study . channel as found from a full multichannel calculation , as a function of energy for different values of magnetic field b. the crossing of the @xmath52 channel threshold is marked by a finite jump in the value of @xmath64 . ] varies linearly with @xmath65 . this agrees with the wigner threshold law . as for the energy dependence of @xmath58 , it can be assumed to be constant and approximated by its value at zero energy . ] to specify our model , we need to find the parameters that characterize the feshbach resonance , @xmath66 and @xmath58 . this is most easily done by fitting the data to the derivative of the phase shift @xmath67 with respect to @xmath47 , as determined by eq . ( [ feshphase ] ) , @xmath68 the resonance width and the location of the dressed bound state can be obtained by fitting a lorentzian to the numerically calculated values as a function of @xmath8 . in fig . [ gammafit ] we plot the values obtained from this procedure . the fit to the data agrees well with the expected wigner threshold law , @xmath69 , where we find @xmath70 @xmath71 . we neglect the small value of the zero intercept , typically associated with the presence of the inelastic component . finally , from fig . [ gammafit ] we see that for @xmath72 @xmath12k , the level shift acquires a constant value of @xmath73 @xmath12k . thus , we model the scattering phase shift by @xmath74.\ ] ] a more complete model at higher energies requires a determination of the threshold law for @xmath58 . _ @xcite have shown for a one channel model or more ( if losses are neglected ) that this law can be obtained from the below - threshold behaviour of the feshbach bound state as a function of the @xmath47 field . while this is true , a direct application of their model is not possible here since we are dealing with a situation where one of the closed channel gets opened . as we are mainly interested in near - threshold behaviour in order to extract information below threshold , we will neglect the functional behaviour of @xmath58 at higher energies . the circles represent values of @xmath75 obtained from multichannel code for @xmath76 , @xmath77 , and @xmath78 gauss . the solid lines correspond to the analytical model developed in this paper . the fit is chosen to be made in the regime where the \|bp > channel is closed . ] channel to negative energies for @xmath76 gauss . there is very good agreement between values obtained using the coupled channel code ( shown by diamonds ) and the analytic model ( solid line ) . ] as in the case of @xmath5 channel , the background scattering length acts as a free parameter that is determined by fitting the above model to one low - energy data point . in fig . [ phasefitfeshbach ] we plot the scattering phase shift @xmath79 for three different values of the magnetic field @xmath47 . the plots show a very good agreement of the analytical model to the numerical multichannel data . given our model , it is simple to obtain the scattering properties such as the @xmath2-wave scattering length at negative energy ( below threshold ) , by analytically continuing the above formula to the positive imaginary axis in the complex @xmath26-plane . figure [ scattfit ] shows excellent agreement of the scattering lengths obtained using our model with the numerical data for positive energies . the predicted values at negative energies are also shown which agree well with the numerical value obtained just above threshold . , plotted as a function of energy . a fit of the analytic inelastic model eq . ( [ iemodel ] ) is shown by solid line to the coupled channel numerical shown by red circles . the dashed line corresponds to the prediction for negative energies based on analytic continuation of the model . ] critical to the use of the tir for quantum coherent coupling between atoms is a favorable ratio of elastic to inelastic scattering processes . near threshold scattering at energies @xmath80 can couple the @xmath37 channel to one other open channel , @xmath41 . such transitions will lead to loss of the atomic pair , as the hyperfine energy will be converted to kinetic energies that are much larger than the depth of the trap . by unitarity of the @xmath0-matrix , such inelastic processes imply @xmath81 , which can be modeled by an imaginary part of the scattering phase shift . formal theory of multichannel threshold scattering allows us to model the energy dependence of this imaginary phase shift @xcite . nesbet has shown how to separate the threshold behaviour arising from smooth background and singularities of poles such as those resulting from virtual bound states and feshbach resonances below threshold @xcite . for a general multichannel situation consisting of @xmath82 open and @xmath83 closed channels below and above a certain threshold , the off - diagonal elements of the @xmath0-matrix near threshold can be written as @xmath84 where @xmath85 is some @xmath86 matrix , @xmath87 , @xmath88 , and @xmath89 is the location of the pole on the imaginary @xmath26 axis . in our case , for @xmath44 , near the @xmath37 channel threshold , this corresponds to one open channel @xmath41 and one closed channel @xmath37 , the relevant s - matrix element has the form @xmath90 where @xmath91 is some proportionality constant that can be treated as a parameter . the imaginary part of the background phase shift is given by the wigner threshold law @xmath92 where @xmath93 is the imaginary part of the background scattering length . the values of the parameters of the model are typically obtained by fitting eq . ( [ iemodel ] ) to the numerical coupled channels data once the position of the pole , @xmath89 , is known . @xcite showed that the existence of a virtual bound state leads to a rapid change in the shift in the energy of a feshbach resonance , @xmath94 in eq . ( [ deriv ] ) , and derived a simple formula that connects the pole location to the shift . from inset of fig . [ boundstate ] , the rapid change in @xmath94 near threshold strongly indicates the existence of a such a virtual pole . we can not , however , apply the formula in @xcite , since that analysis was limited to a two - channel model . here , we have a more complicated interplay of three channels , ( @xmath37 , @xmath41 , @xmath40 ) , since the feshbach resonance arises from the bound state in the @xmath40 channel . a more detailed analysis is required to determine the location of @xmath89 from first principles . instead , since sufficient data points are available from the coupled channels numerical solution , we treat @xmath89 as an additional fitting parameter . obtained by analytic continuation of the multichannel @xmath0-matrix at @xmath95 g. solid lines correspond to numerical coupled channels data while dashed line is the analytic continuation . ] figure [ lossmodel ] shows excellent fit of the above model to the numerical data . analytic continuation of this model will give a precise determination of the off - diagonal elements of the @xmath0-matrix at negative energies . this is shown in fig . [ realandimaglength ] for the magnetic field value @xmath96 g. since the above model for the inelastic component of the @xmath0-matrix is known to be valid only in a small energy region close to threshold , we have limited ourselves to the energy range @xmath97 to @xmath98 for purpose of illustration . as a measure of the ratio of inelastic to elastic processes , we define a general energy - dependent complex scattering length , @xmath99+i \text{im}[a ] = -\tan[\delta(e)]/k,\ ] ] which can be analytically continued to negative energies . in fig . [ realandimaglength ] . we plot @xmath100/\text{re}[a]$ ] . we see that inelastic processes should not dominate , even at negative energies sufficiently far from the feshbach resonance . in this paper we have discussed the need for an analytic model of the @xmath0-matrix in order to extract scattering properties at negative energies where numerical methods fail . such negative energy solutions are essential for understanding trap - induced resonances that involve atoms tunneling into regions of the molecular potential that are below the threshold . beyond improved numerical solutions , these models give us physical intuition with regards to the scattering resonances that are critical to developing many - body model hamiltonians that can help to explain ultra - cold atomic phenomena @xcite . we applied this method to study the collision of two @xmath1cs atoms in separate harmonic traps , a situation similar to that of atoms in separate sites of a polarization gradient optical lattice . colliding atoms in spin states that are chosen so that they are robust with respect to trapping inhomogeneities are not necessarily optimal when considering the scattering process ; they can undergo multichannel scattering processes , including inelastic loss . we studied a specific example of such scattering collisions between @xmath101 and @xmath102 . a feshbach resonance occurs at very small magnetic fields ( @xmath103 30 mg ) due to coupling to a bound state in the @xmath104 channel . to treat this , we modeled the @xmath0-matrix element via a smooth background component with an imaginary part in its scattering length and elastic single - resonance feshbach model . the resulting total scattering length agrees well with direct numerical multichannel scattering solution at small positive energies and extends analytically to negative energies well beyond the validity of numerical solutions . we have also discussed a model consisting of a few parameters in order to describe the off - diagonal element of @xmath0-matrix that corresponds to inelastic processes . the model agrees extremely well with the coupled channels calculation at positive energies above threshold . analytic continuation allows us to calculate the scattering length at negative energies below threshold . we note that , whereas for free space scattering the reaction rates for elastic and inelastic processes just above threshold are simply related to the complex scattering length @xcite , in the negative energy case one must account for the tunneling rate from trap to the molecular potential . we will treat this in detail in a future publication . within the framework described in this paper , it is also possible to study the threshold behaviour arising due to the opening / closing of the @xmath40 channel . this particular threshold is interesting because the channel closing can be controlled by use of a very small magnetic field . also since @xmath40 is the same channel that has the feshbach bound state , opening of this channel leads to the disappearance of this feshbach resonance . while feshbach resonances in ultra - cold atomic gases have been thoroughly studied in recent years , our study opens up the prospect of studying feshbach resonances in the vicinity of such tunable thresholds and their implications to the many - body properties of trapped ultra - cold gases . we are extremely grateful to paul julienne for his hospitality during our visit to nist in relation to this project and for his direction in the operation of the mies - julienne - sando nist close - coupling codes , used to perform the multichannel scattering calculations presented here . we also thank rene stock for helpful discussions on the trap - induced resonances . sb and ihd acknowledge financial support from the onr , grant no . n00014 - 03 - 1 - 0508 , and dto grant no . daad19 - 13-r-0011 . sb also acknowledges financial support from the w. m. keck program in quantum materials at rice university . sk acknowledges financial support from the nwo . 99 c. a. regal , c. ticknor , j. l. bohn , and d. s. jin , nature * 424 * , 47 ( 2003 ) ; m. greiner , c. a. regal , and d. s. jin , nature * 426 * , 537 ( 2003 ) . m. w. zwierlein , c. a. stan , c. h. schunck , s. m. f. raupach , s. gupta , z. hadzibabic , and w. ketterle , phys . lett . * 91 * , 250401 ( 2003 ) ; k. xu , t. mukaiyama , j. r. abo - shaeer , j. k. chin , d. miller , and w. ketterle , phys . lett . * 91 * , 210402 ( 2003 ) . s. jochim , m. bartenstein , a. altmeyer , g. hendl , s. riedl , c. chin , j. hecker denschlag , and r. grimm , science * 302 * , 2101 ( 2003 ) ; s. jochim , m. bartenstein , a. altmeyer , g. hendl , c. chin , j. h. denschlag , and r. grimm , phys . 91 * , 240402 ( 2003 ) . k. e. strecker , g. b. partridge and r. g. hulet , phys . * 91 * , 080406 ( 2003 ) . d. jaksch and p. zoller , ann . phys . * 315 * , 52 ( 2005 ) ; immanuel bloch j. phys . * 38 * s629 ( 2005 ) ; m. lewenstein em et al . , advances in physics , * 56 * , 243 ( 2007 ) , and references therein . s. g. bhongale , m. r. goosen , s. j. j. m. f. kokkelmans , arxiv:0710.5288 ; s. g. bhongale , j. n. milstein , m. j. holland , phys . a * 69 * , 053603 ( 2004 ) . u. dorner , t. calarco , p. zoller , a. browaeys , and p. grangier , j. opt . b * 7 * , s341 ( 2005 ) . p. s. jessen and i. h. deutsch , adv . at . . phys . * 36 * , 91 ( 1996 ) ; and references therein . c. chin et al , science , * 305 * , 1128 ( 2004 ) ; c. a. regal , m. greiner , and d. s. jin , phys . rev . lett . * 92 * , 040403 ( 2004 ) ; m. w. zwierlein et al . , cond - mat/0403049 ( 2004 ) ; t. bourdel et al , phys . * 93 * , 050401 ( 2004 ) ; k.m . o hara et al , science * 298 * , 2179 ( 2002 ) ; r. hulet , presentation in the quantum gas conference at the kavli institute for theoretical physics , santa barbara , ca , may 10 - 14 ( 2004 ) . t. bergeman , m. g. moore , and m. olshanii , phys . lett . * 91 * , 163201 ( 2003 ) . moritz _ rev . lett . * 94 * , 210401 ( 2005 ) . k. winkler _ _ , nature * 441 * , 853 ( 2006 ) . d. b. m. dickerscheid em et al . a * 71 * , 043604 ( 2005 ) . m grupp _ et al . _ j. phys . b : at . mol . . phys . * 40 * 2703 ( 2007 ) . n. nygaard , r. piil , and k. moelmer , unpublished , e - print arxiv:0708:2329 . r. stock , i. h. deutsch , and e. bolda , phys . lett . * 91 * , 183201 ( 2003 ) r. stock and i. h. deutsch , phys . a * 73 * , 032701 ( 2006 ) . p. s. julienne and bo gao , atomic physics * 20 * , proceedings of the 20th icap , innsbruck , austria ( 2006 ) , e - print arxiv : physics/0609013v1 . b. marcelis , e. g. m. van kempen , b. j. verhaar , and s. j. j. m. f. kokkelmans , phys . rev . a * 70 * , 012701 ( 2004 ) . m. l. chiofalo , s. j. kokkelmans , j. n. milstein , and m. j. holland , phys . 88 , 090402 ( 2002 ) . brennen _ et al . lett . * 82 * , 1060 ( 1999 ) . lett . * 82 * , 1975 ( 1999 ) . m. greiner _ nature * 425 * , 937 ( 2003 ) . p. j. leo , c. j. willaims , and p. s. julienne , phys . lett . * 85 * , 2721 ( 2000 ) . e. tiesinga , e. l. bolda , and p. s. julienne , phys . a * 66 * , 013403 ( 2002 ) . c. j. joachain in _ quantum collision theory _ , north holland , third edition(1983 ) . i. h. deutsch and p. s. jessen , phys . a * 57 * , 1972 ( 1998 ) . j. r. taylor in _ scattering theory : the quantum theory of nonrealtivistic collisions _ , john wiley and sons , new york .	we construct simple analytic models of the @xmath0-matrix , accounting for both scattering resonances and smooth background contributions for collisions that occur below the s - wave threshold . such models are important for studying confinement - induced resonances such as those occurring in cold collisions of @xmath1cs atoms in separated sites of a polarization - gradient optical lattice . because these resonances occur at negative energy with respect to the @xmath2-wave threshold , they can not be studied easily using direct numerical solutions of the schrdinger equation . using our analytic model , we extend previous studies of negative - energy scattering to the multichannel case , accounting for the interplay of feshbach resonances , large background scattering lengths , and inelastic processes .
in recent years , a renewed interest in the casimir effect @xcite emerged , as a consequence of new , refined experimental techniques @xcite , which raised the required standards of the theoretical calculations . they were usually based on gross simplifying assumptions , mostly about the properties of the mirrors . to explain the experimental results , it is then important to use more accurate models , including corrections due , for example , to the imperfect nature of the mirrors . in this paper we are concerned with the calculation of the fermionic casimir effect for imperfect ` mirrors ' , in a functional integral framework . in a fermionic context , the mirrors are meant to partially reflect the fermionic current ; including as a particular case the ` bag ' boundary conditions , where the current is completely reflected . from a technical point of view , we shall use an adapted version of a previously used functional approach @xcite , to cope with the case of fermionic fields and imperfect boundaries . in the fermionic context , this method had already been applied to the case of fermions in @xmath3 dimensions satisfying bag conditions on a curve and coupled to a gauge field @xcite . to cope with imperfect mirrors , which will occupy parallel planes , we follow @xcite to introduce a coupling of the bilinear @xmath4 to a @xmath0-function potential of the proper coordinate . we shall then calculate the resulting casimir energy as a function of the distance between two mirrors , using to that effect a functional formalism which maps the original problem to a one - dimensional one , where the fields live on the surfaces of the mirrors . we show that exact results may be obtained by for @xmath1 and @xmath2 dimensions ; for the @xmath1 dimensional case our result in the perfect bag case agrees with the one of @xcite , in spite of the fact that the treatment of the problem is quite different ; in particular , the description of the interaction can not be mapped in a straightforward manner from one calculation to the other . this article is organized as follows : in section [ sec : method ] we introduce the method in general ; in [ sec : casimir ] we apply it to the calculation of the casimir energy for two imperfect mirrors in @xmath1 and @xmath2 spacetime dimensions . in section [ sec : bag ] we analyze the relationship between this kind of interaction and the bag boundary conditions , which emerge when the coupling constant takes a particular value . finally , in section [ sec : conc ] we present our conclusions . in this section , we shall introduce the method used for the calculation of the vacuum energy , in the presence of ` defects ' which generate the approximate bag - like boundary conditions , for the case of a fermionic field in @xmath5 spacetime dimensions . our approach is based on the well - known property that the vacuum energy , @xmath6 , may be obtained from the euclidean functional @xmath7 , as : @xmath8 where @xmath9 is the extension of the ( imaginary ) time coordinate . on the other hand , in the functional integral representation , @xmath10 is given , for a dirac field , by : @xmath11 where @xmath12 , the euclidean action , will be such that @xmath13 , with @xmath14 which are the free and interaction actions , respectively . in the last equation , @xmath15 denotes a ` potential ' , introduced in order to simulate the boundary conditions . although we shall mostly consider just the cases of one or two defects , we write ( for the sake of generality ) its expression for a system with @xmath16 equally - spaced defects : @xmath17 where @xmath18 and @xmath19 . @xmath20 is the coupling constant , whose role in imposing an approximate boundary condition is discussed in section [ sec : bag ] . it is worth noting here that , as mentioned in @xcite , dirac s equation with this kind of potential is not well - defined . however , also in @xcite , a consistent calculation of the casimir energy for a dirac field in @xmath1 dimensions could still be performed , since one can set up the problem in terms that avoid the calculation of the eigenstates of the dirac hamiltonian . we shall here use an approach that also bypasses the eigenproblem for the dirac hamiltonian , relying instead on the corresponding euclidean propagator . the reason this can be done is , of course , that the vacuum energy may be expressed as a function of the expectation values of fermionic bilinears which are , at least in this case , well - defined even when the eigensystem is not . of course , the total energy for this kind of potential will be , since there is translation invariance along the coordinates parallel to the surface of the mirrors , proportional to the ` area ' of each defect ; thus we will be interested in evaluating energy densities , or pressures , rather than the extensive quantity @xmath6 ( except the @xmath21 case where the two quantities coincide ) . in @xmath22 spatial dimensions , the defects are co - dimension @xmath23 hyperplanes . putting the system in a @xmath24-dimensional ` box ' of side @xmath25 , we obtain the vacuum energy density @xmath26 dividing @xmath6 ( calculated for such a box ) by @xmath27 , and then taking the @xmath28 limit : @xmath29 we have denoted by @xmath30 the coordinates in @xmath5 dimensional spacetime @xmath31 . it will turn out to be convenient to introduce also coordinates @xmath32 , for the subset of the @xmath33 spacetime components that parametrize the @xmath34 defect s world - volume , as @xmath35 . analogously , momentum - space coordinates in @xmath5 dimensions are denoted by @xmath36 , while @xmath37 is reserved for its @xmath33-dimensional restriction : @xmath38 where @xmath39 is the momentum component along the normal direction to the mirrors . in terms of these conventions , we have : @xmath40 which , following the methods of @xcite , may be written in an equivalent way by introducing a pair of auxiliary grassmann fields @xmath41 and @xmath42 for each defect . the vacuum functional then adopts the form : @xmath43 \;\exp \big [ \sum_{\alpha = 0}^{n - 1 } \int d^{d+1}y\;\chi^2_\alpha(y ) \nonumber\\ & + & g \ , \int d^{d+1}x\int d^{d+1}x'\ , \bar\eta ( x)\ , g^{(0)}(x , x')\,\eta(x')\big]\,,\end{aligned}\ ] ] where : @xmath44 while @xmath45 is the vacuum functional in the absence of defects : @xmath46 on the other hand , @xmath47 is the free fermion propagator . note that the auxiliary fields are naturally defined on each one of the mirrors world - volumes . since we are interested in calculating the dependence of @xmath26 on the positions , @xmath48 , of the defects , we shall discard any constant which is independent of those variables . constant factors in the functional integral , like @xmath45 , generate precisely that kind of constant , which we shall therefore ignore ( without bothering to rename the corresponding subtracted quantity ) . replacing ( [ def : eta ] ) into ( [ eq : z2 ] ) , we may write the latter in the more explicit form : @xmath49 \ ; e^{\sum_{\alpha,\beta = 0}^{n - 1}\int d^{d+1}y\int d^{d+1}y'\ , \bar\chi_\alpha ( y)\ , \mathcal{k}_{\alpha \beta}(y , y')\,\chi_\beta(y ' ) } \;,\ ] ] where we introduced the matrix kernel : @xmath50 where each @xmath51 element is also a matrix in dirac space . the formal result of evaluating the integral above is then : @xmath52\ ] ] where the determinant is meant to affect all , continuum and discrete , variables . this means that : @xmath53 where ` @xmath54 ' is the trace over continuous and discrete indices . since there is invariance under translations in the @xmath55 coordinates , we may fourier transform with respect to them . the kernel becomes then block diagonal in the continuous variables ; besides , a proper counting of the modes in the box shows that the @xmath56 factor is cancelled by an identical one coming from the numerator ; thus , in the limit , @xmath57 \;,\ ] ] where the tilde has been used to denote fourier transformation , and ` @xmath58 ' denotes trace over the @xmath59 and dirac indices only . in the following section , we deal with the evaluation of the previous expression for @xmath26 in the most relevant case , i.e. , @xmath60 . besides the subtraction of the vacuum energy in the absence of the defects , there is another constant to get rid of : the self - energy of the defects . it may be identified as the result of evaluating @xmath26 for @xmath61 . thus the prescrition for such a subtraction , which will render a finite quantity as a result , is to subtract from @xmath26 its limit when @xmath62 . of course , to give meaning to the expression at intermediate steps , a regularization is introduced . since the last subtraction brings into play the self - energy of the defects , we present first @xmath63 , where the self - energy first emerges , and then the @xmath60 case where we evaluate the energy for the case of two defects , identifying and subtracting the contributions of the self - energies of the two mirrors . for @xmath64 , the @xmath65 matrix has only one element ( @xmath66 ) , thus there is no need to use @xmath67 , @xmath68 indices : @xmath69 so that @xmath26 becomes : @xmath70\big\ } \;.\ ] ] when @xmath21 , the euclidean gamma matrices have the single eigenvalues @xmath71 . then the vacuum persistence amplitude is : @xmath72 which has a logarithmic uv divergence , easily regularized by using a frequency cutoff . a lengthier , but quite straightforward calculation shows that , in @xmath73 : @xmath74\;,\ ] ] which is quadratically divergent . again , an euclidean can be used to give meaning to the integral , whose explicit form we do not need : indeed , it will emerge in the next subsection only to be subtracted in order to fix the energy to zero when @xmath62 . when @xmath75 the matrix @xmath76 is given by @xmath77\ ] ] with @xmath78 as defined in ( [ eq : defw ] ) . using the property : @xmath79\ ] ] for the determinant of the matrix in ( [ eq : k2 ] ) , we may identify the first two factors above as yielding the self - energies for the defects , as studied in [ ssec : one ] . we first introduce an uv cutoff to give meaning to those two factors ( the third factor is convergent ) before subtracting the corresponding self - energies . the resulting pressure is : @xmath80 \;,\ ] ] with @xmath81 let us now evaluate @xmath26 above for @xmath21 and @xmath82 . in one spatial dimension , if we take into account dirac s indices , @xmath83 is a matrix with two @xmath84 blocks , while @xmath85 is just a @xmath86 matrix . a somewhat lengthy but otherwise straightforward calculation shows that the latter has eigenvalues @xmath87 given by : @xmath88 ^ 2}\,e^{-2l\|q_0\| } \;,\end{aligned}\ ] ] where @xmath89 . therefore , the trace in the expression for the vacuum energy can be calculated exactly . in the massless case , we can obtain the exact expression for the energy as a function of @xmath20 , since : @xmath90\;,\ ] ] or , @xmath91 where @xmath92\,,\ ] ] is a dimensionless constant which determines the strength of the casimir interaction , and @xmath93 denotes the polylogarithm function . in figure ( [ fg : casimir2pd1 ] ) , we plot the dimensionless combination @xmath94 as a function of the ( also dimensionless ) variable @xmath20 , for the massless case , and for different values of @xmath95 , corresponding to the massive case ( where there is no closed expression for the energy ) . note that there is a maximum at @xmath96 . that value , as explained in the next section , corresponds to exact bag boundary conditions , where the boundaries are more effective and as a consequence the energy reaches it maximum possible value : @xmath97_{g=2 } \;=\ ; - \frac{\pi}{24 l}\;\ ] ] note that in the approach of reference @xcite , bag conditions correspond to an infinite coupling constant . the difference is explained , of course , by the different meaning of the coupling constants in both cases . for example , the propagator in our approach includes tadpole - like propagation between the wall and itself , which naturally change the effective value of the coupling constant . nevertheless , our @xmath96 agrees with @xmath98 . a fuller explanation of this fact is given in section [ sec : bag ] , where we derive the form of the fermion propagator in the presence of the defects , studying its behaviour close to the boundary . ( 0,0 ) as a function of @xmath20 , for different values of @xmath99.,title="fig : " ] # 1#2#3#4#5 @font ( 4662,3001)(1414,-3437 ) ( 1441,-691)(0,0)[lb ] as already mentioned , for the massive case the corresponding integral can not be performed exactly . in figure 2 we plot the result of numerically evaluating that integral for @xmath100 as a function of @xmath99 , for different values of @xmath20 . ( 0,0 ) as a function of @xmath101 ; @xmath102.,title="fig : " ] # 1#2#3#4#5 @font ( 5683,3053)(628,-2911 ) ( 643,-161)(0,0)[lb ] in @xmath103 spatial dimensions , each block in @xmath104 is a @xmath105 matrix . for a massive field , and using the notation @xmath106 we find that : @xmath107 ^ 2 } \big [ - 2\,q^2 \,{\mathcal i}\,+\ , g\,q^2\,\gamma_d\nonumber\\ & + & i\,(2\,m\,+\,g\,\omega(q))\,\not \ ! q \,+\,i\ , ( 2\,\omega(q)\,+ g\,m)\,{\not \ ! q}\,\gamma_d \big]\;,\end{aligned}\ ] ] with @xmath108 , @xmath109 and @xmath110 . the exact eigenvalues of @xmath104 can be found after some algebra , the result being : @xmath111 ^ 2}\,e^{-2 \omega(q ) l } \;,\end{aligned}\ ] ] each one with multiplicity equal to two . then , after integrating out the angular variables , the casimir energy per unit area may be written as follows : @xmath112 where @xmath113 ^ 2}\big\}\;,\ ] ] which converges in both the ir and uv regions . it is convenient to introduce first the result corresponding to the case @xmath114 and @xmath96 , where the energy reaches a maximum : @xmath115 in agreement with @xcite ( see also @xcite ) . when @xmath116 , and introducing @xmath117 , @xmath118 ^ 2}\big ] \;.\ ] ] this integral can not be found analytically , but it has a convenient form for performing it numerically . in [ fg : evsg.jpg ] , we plot @xmath119 as a function of @xmath20 for different values of @xmath120 between @xmath121 and @xmath23 . we see the expected decreasing behaviour of the energy when @xmath122 grows , regardless of the value of @xmath20 . this happens because the fermionic field propagator has a faster decay with distance the larger is the mass , reducing the interaction energy between them . on the other hand , the energy always has a maximum when @xmath123 , for any value of the mass . as we mentioned for the @xmath21 case , this comes from the bag boundary conditions , which are met precisely at this value . ( 0,0 ) as a function of @xmath20 , for different values of @xmath124.,title="fig : " ] # 1#2#3#4#5 @font ( 4908,3091)(346,-3107 ) ( 451,-286)(0,0)[lb ] in figure [ fg : evsm.jpg ] , we present the complementary view , by plotting @xmath125 as a function of @xmath124 , for different values of @xmath20 between @xmath121 to @xmath103 . it can be seen that the energy becomes negligible when @xmath126 which means that the distance between the plates is of the order of the decay length of the propagator . ( 0,0 ) as a function of @xmath99 , for different values of @xmath20.,title="fig : " ] # 1#2#3#4#5 @font ( 5390,2870)(414,-2551 ) ( 429,125)(0,0)[lb ] bag boundary conditions are usually formulated in terms of the fermionic propagator , in the presence of boundaries . for the case of just one defect located at @xmath127 , one says that it imposes bag conditions on its two , ` right ' and ` left ' faces , when : @xmath128 respectively . they imply that the vacuum expectation value of the normal component of the fermionic current vanishes when approaching the wall from each side . similar conditions may be imposed , of course , on more than one wall . following an analogous procedure to the one use for the calculation of @xmath10 , but now for the propagator : @xmath129 we obtain : @xmath130 where : @xmath131_{\alpha\,\beta}(q)\,\tilde{g^{(0)}}(q , a_\beta , x'_d)\,.\end{aligned}\ ] ] with @xmath132 the matrix elements of the inverse of @xmath133 may be written in terms of the matrix elements of @xmath133 : @xmath134^{-1 } \tilde{\mathcal k}_{01 } c_1^{-1}}\\ \\ { -c_1^{-1 } \tilde{\mathcal k}_{10 } [ \tilde{\mathcal k}_{00}]^{-1 } } & { c_1^{-1 } } \end{array } \right ) \;,\ ] ] where @xmath135^{-1}$ ] and @xmath136^{-1}$ ] denote the inverses of the respective matrix elements ( not to be confused with the matrix elements of the inverse ) . we have used the definitions @xmath137^{-1 } \tilde{\mathcal k}_{10}$ ] and @xmath138^{-1 } \tilde{\mathcal k}_{01}$ ] . let us apply the formulae above , for the sake of simplicity , to the case of massless fermions and only one defect in @xmath139 , we see that , when @xmath140 and @xmath141 we obtain : @xmath142 \,=\ , e^{-\,\vert q \vert \vert { x'}_1\vert}\,\big(\frac{2\,g}{4 + g^2}\;\mp\ , \frac{1}{2}\big)\\ ] ] @xmath143\,\;,\ ] ] where @xmath144 is the sign function . since we are interested in the @xmath145 sign when approaching the defect from the right , we see that the bag condition is fulfilled when @xmath146 thus , @xmath96 . on the other hand , when one approaches the defect from the left , one also obtains @xmath96 to satisfy the corresponding bag condition . an entirely analogous derivation yields the same value for @xmath20 when the fermions are massive , or when more dimensions are considered . for example , in @xmath3 dimensions , and for @xmath114 the propagator reduces to the one of @xcite . we derived general expressions for the vacuum energy in the presence of @xmath16 ` defects ' , whose role is to impose _ imperfect _ bag - like boundary conditions on parallel planes . standard bag conditions may be obtained as a particular case . for the case of two plates , we obtain more explicit formulae , which yield the casimir energy as an integral over a single variable . that integral is both uv and ir finite . we show that , at any given distance between the plates , the energy reaches a maximum precisely when the bag condition is satisfied . our way of implementing the calculation relies on the calculation of the exact fermion propagator in the presence of the defects . this propagator is a perfectly well - defined object , in spite of the fact that the eigensystem for the associated dirac hamiltonian is not well - defined . this ambiguity has the consequence of allowing for different parametrizations of the strength of the coupling between the dirac field and the mirrors . in particular , our method produces a fermionic propagator which satisfies bag boundary conditions when the coupling constant @xmath96 . that value corresponds , in the setting of reference @xcite to an infinite coupling constant . c. d. f. and e. l. l. acknowledge support from conicet , anpcyt and uncuyo . bib g. plunien , b. mller , and w. greiner , phys . rep . * 134 * , 87 ( 1986 ) ; p. milonni , _ the quantum vacuum _ ( academic press , san diego , 1994 ) ; v. m. mostepanenko and n. n. trunov , _ the casimir effect and its applications _ ( clarendon , london , 1997 ) ; m. bordag , _ the casimir effect 50 years later _ ( world scientific , singapore , 1999 ) ; m. bordag , u. mohideen , and v. m. mostepanenko , phys . rep . * 353 * , 1 ( 2001 ) ; k. a. milton , _ the casimir effect : physical manifestations of the zero - point energy _ ( world scientific , singapore , 2001 ) ; s. reynaud _ et al . _ , c. r. acad . paris * iv-2 * , 1287 ( 2001 ) ; k. a. milton , j. phys . a : math . gen . * 37 * , r209 ( 2004 ) ; s.k . lamoreaux , rep . prog . phys . * 68 * , 201 ( 2005 ) ; special issue _ `` focus on casimir forces '' _ , new j. phys . * 8 * ( 2006 ) . lamoreaux , phys . 78 * , 5 ( 1997 ) ; u. mohideen and a. roy , phys . lett . * 81 * , 4549 ( 1998 ) ; b.w . harris , f. chen , and u. mohideen , phys . a * 62 * , 052109 ( 2000 ) ; t. ederth , phys . a * 62 * , 062104 ( 2000 ) ; h.b . chan , v.a . aksyuk , r.n . kleiman , d.j . bishop , and f. capasso , science * 291 * , 1941 ( 2001 ) ; h. b. chan , v.a . aksyuk , r.n . kleiman , d.j . bishop , and f. capasso , phys . lett . * 87 * , 211801 ( 2001 ) ; g. bressi , g. carugno , r. onofrio , and g. ruoso , phys . lett . * 88 * , 041804 ( 2002 ) ; d. iannuzzi , i. gelfand , m. lisanti , and f. capasso , proc . usa * 101 * , 4019 ( 2004 ) ; r.s . decca , d. lpez , e. fischbach , and d.e . krause , phys . lett . * 91 * , 050402 ( 2003 ) ; r.s . * 94 * , 240401 ( 2005 ) ; r.s . decca _ et al . _ , annals of physics * 318 * , 37 ( 2005 ) . m. bordag , d. robaschik , and e.r . wieczorek , annals of physics * 165 * , 192 ( 1985 ) ; r. golestanian and m. kardar , phys . rev . * a 58 * , 1713 ( 1998 ) ; c. d. fosco , f. c. lombardo , and f. d. mazzitelli , phys . d * 76 * , 085007 ( 2007 ) . c. d. fosco and f. d. mazzitelli , phys . d * 74 * , 025020 ( 2006 ) [ arxiv : hep - th/0606092 ] . p. sundberg and r. l. jaffe , annals phys . * 309 * , 442 ( 2004 ) [ arxiv : hep - th/0308010 ] . k. johnson , acta phys . b * 6 * , 865 ( 1975 ) . h. queiroz , j. c. da silva , f. c. khanna , j. m. c. malbouisson , m. revzen and a. e. santana , annals phys . * 317 * , 220 ( 2005 ) [ erratum - ibid . * 321 * , 1274 ( 2006 ) ] [ arxiv : hep - th/0411228 ] . c. d. fosco , a. p. c. malbouisson and i. roditi , phys . b * 609 * , 430 ( 2005 ) [ arxiv : hep - th/0412229 ] .	we use a functional approach to calculate the casimir energy due to dirac fields in interaction with thin , flat , parallel walls , which implement _ imperfect _ bag - like boundary conditions . these are simulated by the introduction of @xmath0-like interactions with the walls . we show that , with a proper choice for the corresponding coupling constants , bag - model boundary condition are properly implemented . we obtain explicit expressions for the energies in @xmath1 and @xmath2 dimensions , for massless and massive fields .
the general problem of externally generated time harmonic waves linearly interacting with a structure of limited spatial extent is basic in many fields of science @xcite . in recent years much work has been done elucidating the consequences for the scattering of waves in cases in which , in the geometric optics approximation , the ray orbits within the structure are chaotic . examples include optical @xcite , acoustic @xcite , microwave @xcite and electronic cavities @xcite . in the case of complex or irregularly shaped enclosures that are large compared with a wavelength , small changes in the frequency and the configuration give rise to large changes in the scattering characteristics . this feature motivates treatments that are statistical in nature . in this regard random matrix theory @xcite has proven useful in predicting universal aspect of chaotic wave scattering problems in the cases of both time reversal symmetric systems ( corresponding to matrix statistics of the gaussian orthogonal ensemble , goe ) and time reversal symmetry broken systems ( corresponding to matrix statistics of the gaussian unitary ensemble , gue ) . scattering problems can be characterized by the scattering matrix @xmath0 which relates outgoing scattered wave amplitudes @xmath7 to incoming waves @xmath8 , via @xmath9 . an alternative formulation is in terms of the impedance matrix @xmath1 . to illustrate the impedance description , consider an electromagnetic wave scattering problem in which @xmath10 transmission lines labeled @xmath11 of characteristic impedance @xmath12 are connected to a cavity . let @xmath13 and @xmath14 represent the voltage and current on transmission line @xmath15 as measured at a suitable reference plane . then the incident wave @xmath16 and the reflected wave @xmath17 may be expressed as @xmath18 , @xmath19 . the impedance matrix @xmath1 relates the vector voltage to the vector current , via @xmath20 , and @xmath1 and @xmath0 are related by @xmath21 , where @xmath22 is the @xmath10-dimensional identity matrix , and @xmath23 . the impedance formulation is identical to the so called r - matrix " , a formulation introduced by wigner and eisenbud in nuclear - reaction theory in 1947 , and further developed in refs . @xcite . statistical variations of the elements of @xmath1 and @xmath0 due to small random variations in the scattering are of great interest . these statistics have two fundamental influences , ( i ) universal aspects described by random matrix theory , and ( ii ) nonuniversal aspects dependent upon the details of the coupling of input channels ( e.g. , transmission lines ) to the scatterer . our main result concerns the quantity , @xmath24}{\sqrt{var[z_{ii}]var[z_{jj } ] } } , \qquad i\neq j , \label{eq : vrz}\ ] ] where @xmath25 $ ] , the variance of the complex scalar @xmath26 , is defined as the sum of @xmath27 $ ] and @xmath28 $ ] . our result is of the form @xmath29 where @xmath30 is a parameter characterizing the losses within the scatterer . for example , in the case of an electromagnetic cavity , @xmath31 , where @xmath32 is the frequency of the incoming signal , @xmath33 is the average spacing between cavity resonant frequencies near @xmath32 , and @xmath34 is the quality factor of the cavity ( @xmath35 if there are no internal losses ) . the remarkable aspect of ( [ eq : vrz_general ] ) is that @xmath36 depends only on the loss parameter and not on the nonuniversal properties of the coupling to the cavity . thus @xmath2 is a universal function of the loss @xmath30 . the results for @xmath37 and @xmath38 ( to be derived subsequently ) are shown in fig . [ fig : vrz ] . for @xmath39 , @xmath40 for @xmath41 , @xmath42 versus the loss parameter @xmath30 , as specified in eq . ( [ eq : vrz_md ] ) and eq . ( [ eq : vrz_md2 ] ) . ] a ratio similar to ( [ eq : vrz ] ) can also be considered for the scattering matrix @xmath0 , @xmath43}{\sqrt{var[s_{ii } ] var[s_{jj } ] } } , \qquad i\neq j. \label{eq : vrs}\ ] ] in contrast with ( [ eq : vrz_general ] ) , @xmath3 in general depends on _ both _ the coupling to the cavity and on the loss parameter @xmath30 . however , in the special case of high loss , @xmath39 , @xmath3 becomes universal , @xmath44 that is , @xmath45 for @xmath39 . based on their electromagnetics experiments , fiachetti and michelsen @xcite have recently conjectured the universality of ( [ eq : vrs_highloss ] ) in the goe case . more generally , ( [ eq : vrs_highloss ] ) follows from classic results of hauser and feshbach describing fluctuations in the cross section of inelastic neutron scattering @xcite , and this result has been obtained by friedman and mello @xcite using the concept of maximization of information entropy , and by agassi _ _ @xcite using a random - matrix model . the important point is that a universal result for @xmath3 [ i.e. , eq . ( [ eq : vrs_highloss ] ) ] applies only for @xmath39 , while the universal result for @xmath2 , eq . ( [ eq : vrz_general ] ) and fig . [ fig : vrz ] , is for arbitrary @xmath30 . in what follows we derive these results and test them by comparison with data obtained from scattering measurements on a chaotic microwave cavity . section ii derives the results for impedance variance ration , @xmath2 . section iii considers the scattering variance ratio , @xmath3 . section iv presents our experimental tests of the theory . in this section we will obtain the universal functions @xmath36 in eq . ( [ eq : vrz_general ] ) . we adopt a formulation ( e.g. , see refs . @xcite ) that incorporates the nonuniversal effects of the specific coupling geometry of input - output channels to the scatterer , combined with the random matrix theory for the universal aspects of the chaotic wave behavior within the scatterer . ( in what follows , we use terminology appropriate to microwave experiments . ) beginning with the case of zero loss ( @xmath46 ) , we have that , in the goe case , the impedance matrix @xmath1 is described by the following statistical model @xcite , @xmath47 here @xmath48 is the wave number corresponding to the incoming frequency , @xmath49 is a vector whose elements are real , independent , zero mean , gaussian random variables of unit variance . the eigenvalues @xmath50 are randomly chosen with statistics appropriate to goe . this is done by generating ensembles of eigenvalues of random matrices and then scaling them such that the mean spacing between adjacent eigenvalues near @xmath50 is @xmath51 . here @xmath52 is given by the weyl formula for the mean eigenvalue spacing . for example , for the case of a 2d cavity @xmath52 is independent of @xmath48 , @xmath53 where @xmath26 is the area of the cavity . the system - dependent part of the coupling is characterized by the corresponding radiation impedance matrix @xmath54 . note also that , since @xmath55 depends only on local geometry near the entrances to the scattering region , its frequency dependence is much slower than that of @xmath1 which depends on the geometry of the scattering region assumed to be characterized by much longer length scales . the radiation impedance is the impedance seen at the reference plane of the input channel when waves that leave the channel propagate outward and are not reflected back to their input by the distant walls of the cavity . thus the radiation impedance is the impedance seen when the distant walls are removed to infinity , or ( as can be done in an experimental measurement @xcite ) when the distant walls are lined with absorber . hence the radiation impedance depends only on the local structure in the vicinity of the port coupling to the cavity and not on the shape or chaotic properties of the cavity . in the case of ports that are far apart , e.g. , of the order of the cavity size , the off - diagonal elements of @xmath55 are small and will be neglected . thus we will take @xmath55 to be a diagonal matrix with elements @xmath56 . in the gue case , the elements of @xmath49 in ( [ eq : zcav_paper2 ] ) are complex with real and imaginary parts each individually having gaussian statistics , @xmath57 in ( [ eq : zcav_paper2 ] ) is replaced by @xmath58 ( where @xmath59 denotes the conjugate transpose ) , and the statistics of @xmath50 are now those appropriate to gue . equation ( [ eq : zcav_paper2 ] ) is derived in @xcite . it results from a formal series expansion of the solution for @xmath1 in terms of eigenfunctions of the closed cavity . the local structure of the eigenfunctions is then assumed to satisfy the so - called `` random plane wave hypothesis '' ; namely , that in the semi - classical limit ( @xmath60 is large compared to the typical dimensions of the structure ) , if ray trajectories in the closed cavity are chaotic , then eigenfunction @xmath61 is statistically similar to a random isotropic superposition of plane waves of wave number @xmath62 with magnitude@xmath63 . this gives eq . ( [ eq : zcav_paper2 ] ) @xcite . we then further supplement eq . ( [ eq : zcav_paper2 ] ) with the known distribution function of @xmath50 determined from random matrix theory and the local mean spacing @xmath64 between adjacent eigenvalues . the universality of the result for @xmath2 ( fig . 1 ) can be shown from ( [ eq : zcav_paper2 ] ) as follows . the impedance matrix @xmath1 defined in ( [ eq : zcav_paper2 ] ) will have a mean and a fluctuating component . as shown in refs @xcite , the value of the mean is determined by all terms in the sum and thus depends on the slow @xmath48-dependence of @xmath65 and @xmath52 , while the fluctuating component of @xmath1 is determined by terms in the sum in eq . ( [ eq : zcav_paper2 ] ) for which @xmath66 . thus if @xmath67 and @xmath51 are approximately constant over a range of @xmath50 values corresponding to many resonances , then the fluctuating part of @xmath1 will be universal after appropriate normalization . if we define @xmath68 to be the fluctuating part of @xmath1 , we have @xmath69 where @xmath70 are the eigenvalues of a goe random matrix , normalized to have unit mean spacing between adjacent eigenvalues near @xmath71 . thus @xmath72 is a universal normalized impedance matrix . if @xmath55 is diagonal , we have @xmath73 . hence the ratios @xmath2 defined by ( [ eq : vrz ] ) is the same for @xmath1 and @xmath72 , and is , therefore , universal . ( we emphasize , however , that this conclusion relies on @xmath55 being diagonal . ) the effects of distributed loss , such as losses due to conducting walls or a lossy dielectric that fills the cavity , can be simply incorporated in eq . ( [ eq : zcav_paper2 ] ) . since modal fluctuations in losses are small when the modes are chaotic and the wavelength is short , we can construct a complex cavity impedance accounting for distributed loss by simply replacing @xmath74 in eq . ( [ eq : zcav_paper2 ] ) by @xmath75 , where @xmath34 is the cavity quality factor . ( in terms of the normalized impedance matrix @xmath72 defined in ( @xmath76 ) we replace the denominator @xmath70 by @xmath77 $ ] . ) for example , @xmath78 . \end{aligned } \label{eq : z11_goe}\ ] ] ( henceforth , we employ the notation @xmath79 . ) calculation of the moments of the impedance is facilitated by the fact that the eigenvalues and eigenfunctions in the chaotic cavities are statistically independent . for example , the expected value of @xmath80 is , @xmath81&=\lim_{m\rightarrow \infty}\frac{1}{\pi}\sum_{n=1}^m\int dw_{in } f(w_{in } ) w_{in}^2 \int dk_1 ^ 2 \cdots dk_m^2\\ & p_j(k_1 ^ 2 , \cdots , k_m^2 ) \frac{r_{ri}\delta_n ( k_{n}^2-k^2)}{(k^2-k_{n}^2)^2+(k^2/q)^2 } , \end{aligned } \label{eq : e1x11_joint}\ ] ] where @xmath82 is the probability distribution function ( pdf ) of @xmath83 and @xmath84 is the joint pdf of the eigenvalues . integrating over all @xmath85 , @xmath86 , we express @xmath87 $ ] as an integral over the pdf of @xmath50 , @xmath88 , we consider the @xmath89 limit and use @xmath90 for the gaussian random variable @xmath49 . @xmath81 = \int dk_{n}^2 \frac{r_{ri}(k_n)(k_{n}^2-k^2)/\pi } { ( k^2-k_{n}^2)^2+(k^2/q)^2}=x_{ri}(k ) . \end{aligned } \label{eq : e1x11}\ ] ] the second equality in ( [ eq : e1x11 ] ) relating @xmath87 $ ] to the radiation reactance requires @xmath91 and is analogous to the kramers - kronig relation . the second moment of @xmath80 can be determined in a similar way by integrating over all @xmath92 except @xmath93 and using the joint distribution function @xmath94/(m\delta)^2 $ ] , where @xmath95 is known from random matrix theory @xcite . using the fact that @xmath96 goes to zero at large argument and assuming that the radiation resistance @xmath97 and the average spacing @xmath98 vary slowly over the damping width @xmath99 , we obtain @xmath100= \frac{3}{2\pi}(r_{ri}^2\frac{\delta}{k^2/q})+\{e[x_{ii}]\}^2 + \frac{r_{ri}^2}{\pi^2}\int dk_t^2 dk_s^2 \\ & \frac{\langle w^2 \rangle ^2 g(\|k_t^2-k_s^2\|/\delta)(k_t^2-k^2)(k_s^2-k^2)}{[(k_t^2-k^2)^2+(k^2/q)^2 ] [ ( k_s^2-k^2)^2+(k^2/q)^2 ] } , \end{aligned } \label{eq : e2x11}\ ] ] combining eq . ( [ eq : e2x11 ] ) and eq . ( [ eq : e1x11 ] ) , we obtain @xmath101=\frac{r_{ri}^2}{\lambda}[\frac{3}{2\pi}-\frac{1}{\pi } \int_0^{\infty } d x g(x ) \frac{4}{4+(x/\lambda)^2 } ] , \label{eq : var_x11_simp}\ ] ] where @xmath102 . a similar moment evaluation can be carried out for @xmath103 , as specified in eq . ( [ eq : z11_goe ] ) , which yields the same expression as eq . ( [ eq : var_x11_simp ] ) for @xmath104 $ ] . for goe ( the case we are now considering ) we have that @xcite , @xmath105(d f / d s)$ ] , where @xmath106 $ ] . in order to obtain the variance ratio , we also apply the previous process to the off diagonal term @xmath107 , @xmath108 , which , based on eq . ( [ eq : zcav_paper2 ] ) , is given by @xmath109 . \end{aligned } \label{eq : z12}\ ] ] since @xmath83 and @xmath110 are independent , the first moments of @xmath111 and @xmath112 are both zero , and the variance is equal to the second moment , @xmath113&=\lim_{m\rightarrow \infty}\frac{m}{\pi^2}\int dk_{n}^2 \frac{r_{ri}r_{rj}\delta_n^2\langle w_{in}^2\rangle \langle w_{jn}^2\rangle } { [ ( k_{n}^2-k^2)^2+(k^2/q)]^2}p_1(k_{n}^2 ) \\ & = \frac{r_{ri}r_{rj}}{\lambda}\frac{1}{2\pi } , \end{aligned } \label{eq : var_x12}\ ] ] the same result is obtained for @xmath114 $ ] . combining eq . ( [ eq : var_x12 ] ) with eq . ( [ eq : var_x11_simp ] ) , we have eq . ( [ eq : vrz_general ] ) with @xmath115^{-1}. \end{aligned } \label{eq : vrz_md}\ ] ] a similar calculation in the gue case is facilitated by the simpler form of the function @xmath116 which is now given by @xmath117 . we obtain @xmath118^{-1 } \\ & = [ 1+\frac{1-e^{-4 \pi \lambda}}{4 \pi \lambda}]^{-1}. \end{aligned } \label{eq : vrz_md2}\ ] ] we note that subsequent to a previous announcement of our work @xcite , the two - frequency correlation functions for the elements of the impedance and the scattering matrix have recently been calculated by savin , fyodorov and sommers @xcite , and are consistent with the preceding in the limit of zero frequency separation . we now consider the scattering matrix in the high loss limit , @xmath39 . for simplicity , we consider the case of two channels connecting to the scatterer , @xmath119 , and @xmath1 and @xmath0 are @xmath120 matrices . we note that a chaotic scattering process can be divided into a direct process and a delayed process , which leads to a separation of the mean part ( equal to @xmath55 ) and the fluctuating part @xmath121 of the cavity impedance , @xmath122 . the fluctuating part @xmath121 decreases as loss increases . thus in the high loss limit , @xmath123 , which implies @xmath124 . ( recall , the mean parts of the off diagonal components are zero . ) we may now form @xmath0 using @xmath125 . since the off diagonal terms of @xmath1 are small , the diagonal components of @xmath0 are dominated by the diagonal components of @xmath1 . we then find for @xmath126 , @xmath127\delta z_{11 } , \end{aligned } \label{eq : s11dz}\ ] ] where @xmath128 , and @xmath129 is the characteristics impedance of channel 1 . thus , we obtain @xmath130=\|\frac{2z_{01}}{(z_{r1}+z_{01})^2}\|^2 var[z_{11 } ] . \label{eq : var_s11}\ ] ] in addition , we can express @xmath131 in the high damping limit as @xmath132 which leads to @xmath133=\|\frac{2\sqrt{z_{01}z_{02}}}{(z_{r1}+z_{01})(z_{r2}+z_{02})}\|^2 var[z_{12 } ] , \label{eq : var_s12}\ ] ] and similarly for @xmath134 $ ] . combining eq . ( [ eq : var_s11 ] ) and eq . ( [ eq : var_s12 ] ) , we recover eq . ( [ eq : vrs_highloss ] ) and we note that this result is independent of the coupling ( i.e. , independent of @xmath55 ) . to illustrate the influence of coupling on @xmath3 at finite loss parameter @xmath30 , we consider the impedance matrix in the goe case using the model normalized impedance @xmath72 used in ref . @xcite , @xmath135 , where @xmath72 is given by @xmath136 , @xmath137 , and @xmath138 is set to be @xmath139 , such that mean of @xmath72 is zero . realizations of @xmath72 are produced numerically by generating gaussian random variables @xmath83 and spectra @xmath140 from the eigenvalues of random matrices . we express a model scattering matrix @xmath0 as @xmath141 where @xmath142 and @xmath143 , when @xmath144 is the identity matrix and @xmath145 is zero , we reach the so - called perfect coupling condition , which means that the scattering is determined by the delayed process and the direct process is absent . we now consider an example in which the two - port couplings are the same so that @xmath146 , where @xmath147 is a scalar . figures [ fig : nonpcvrs](a ) and ( b ) show results for the variation of @xmath3 with the coupling parameters @xmath148 and @xmath149 , for a high loss case ( @xmath150 ) and for the lossless case ( @xmath46 ) . in fig . [ fig : nonpcvrs](a ) , we fix @xmath151 to be zero , and vary @xmath152 , while in fig . [ fig : nonpcvrs](b ) , @xmath152 is fixed to be 1 and @xmath151 is varied . compared to the high damping case , @xmath3 in the lossless case has a much larger deviation from the constant 1/2 . note that @xmath3 is @xmath153 in the perfect - coupling case ( i.e. , @xmath154 , @xmath155 ) , no matter whether the cavity is highly lossy or lossless . this is related to the concept of weak localization " reviewed in @xcite . versus @xmath152 for @xmath155 in the lossless case @xmath46 and in a high loss case @xmath150 . ( b)@xmath3 versus @xmath151 for @xmath154.,title="fig : " ] versus @xmath152 for @xmath155 in the lossless case @xmath46 and in a high loss case @xmath150 . ( b)@xmath3 versus @xmath151 for @xmath154.,title="fig : " ] in the case of an @xmath10-port we can think of the above two - port consideration of @xmath3 as applying to the @xmath10-port converted to a two port by opening channels @xmath156 ; i.e. , the incoming waves @xmath157 are identically zero ( for a microwave cavity with transmission line inputs , this corresponds to terminating transmission lines @xmath156 with their characteristic impedances , @xmath158 ) . thus ports @xmath156 effectively add to the loss due to the energy flux leaving through them . if the ports @xmath156 are assumed to act like distributed loss , they can be taken into account by increasing the loss parameter @xmath30 . [ this increased loss enhances the validity of eq . ( [ eq : vrs_highloss ] ) . ] we provide experimental results testing the theoretical predictions for the statistical fluctuations in the variance of the @xmath0 and @xmath1 elements , in the limit of large damping . the experiments are done in an air - filled , quarter bow - tie shaped cavity which acts as a two - dimensional resonator below 19.05 ghz ( fig . this cavity has previously been used for the successful study of the eigenvalue spacing statistics @xcite , eigenfunction statistics @xcite , and for studying the universal fluctuations in the impedance @xcite and scattering matrix @xcite for a wave chaotic system . the cavity is driven by two ports ; each of which consists of the center conductor ( diameter 2a=1.27 mm ) of a coaxial cable that extends from the top lid of the cavity and makes contact with the bottom plate of the cavity ( fig . 3(b ) ) . from direct @xmath159 measurements , we estimate that the cavity has a typical loaded @xmath34 of about 150 between 4 - 5ghz and about 300 between 11 - 12ghz . this translates to a damping parameter of @xmath160 for the entire frequency range of this experiment . hence we examine experimentally the time - reversal symmetric ( goe ) cases for the @xmath1 and @xmath0-variance ratios in the high damping limit . to perform an approximation to ensemble averaging , two perturbers ( shown gray in fig . 3(a ) ) , made up of rectangular ferromagnetic solids wrapped in al foil ( of dimensions 26.7 x 40.6 x 7.87 @xmath161 ) , are systematically scanned and rotated throughout the volume of the cavity by means of a magnet that is placed outside the cavity . the ensemble set consists of one - hundred different positions and orientations of the perturbers within the cavity . the perturbers are on the scale of half the wavelength or bigger over the entire frequency range of the experiment . the full two by two @xmath0 matrix is measured between 4 and 12 ghz for each position of the perturbers . below 4 ghz the mode density is too low to obtain meaningful ensemble averaging , while above 12 ghz the coupling becomes too weak to couple to the modes of the cavity , at least for this port geometry . once the @xmath0 matrix has been measured , it is then converted to @xmath1 through @xmath162 . to eliminate the effect on the average of short ray orbits returning to the antenna ( these lead to rapidly varying - with - frequency systematic deviations of the average from the ensemble average , as discussed in @xcite ) we perform frequency averaging over a sliding window of width 300 mhz . we denote such sliding averages of impedance and scattering variance ratios by @xmath163 and @xmath164 . the inset in fig . 4 shows @xmath163 ( solid line ) over a frequency range 4 ghz to 12 ghz . denoting the average of @xmath163 over the entire range , 4ghz-12ghz , by @xmath165 , we obtain @xmath166 , and we find that @xmath167 over the entire frequency range . this value of experimentally obtained @xmath168 is close to the ideal theoretical value of 1/2 for large damping . also shown in the inset is the variance ratio obtained with no frequency averaging ( small circles ) these are deviations from the frequency averaged ratio values with a standard deviation of 0.04 . nevertheless the mean value of the variance ratio over the entire frequency range is 0.49 . the circles , stars and dashes in fig . 4 show the variation in @xmath169 $ ] , @xmath170var[z_{22}]}]$ ] and @xmath171var[z_{22}]}]$ ] respectively , as a function of frequency . the agreement is quite good ( i.e. , the dashes overlie the open circles ) at all frequencies . $ ] ( circles ) , @xmath172var[z_{22}]}$](stars ) and @xmath173var[z_{22}]}$](dashes ) are plotted on a natural - logrithmic scale as a function of frequency from 4 to 12 ghz . inset shows the ratio @xmath163 versus frequency ( solid line ) . the small circles show @xmath2 without any frequency averaging . ] similarly in fig . 5 , we present data for the scattering variance ratio . experimentally we obtain @xmath174 and @xmath175 over the range 4 ghz to 12 ghz . the circles , stars and dashes in fig . 5 show the variation in @xmath134 $ ] , @xmath176var[s_{22}]}$ ] and @xmath177var[s_{22}]}$ ] , respectively , as a function of frequency . similar to the impedance data ( fig . [ fig : samy2 ] ) , we observe that the data for @xmath177var[s_{22}]}$](dashes ) overlie the data for @xmath134$](open circles ) , again indicating that the experimentally obtained value for @xmath178 shows good agreement with the asymptotic theoretical values for highly damped time - reversal symmetric systems over a large frequency range . to sum up , we have used random matrix theory in conjunction with the radiation impedance characterizing the system - dependent coupling details to evaluate the variance ratios ( [ eq : vrz ] ) and ( [ eq : vrs ] ) . the main result is that the impedance variance ratio ( [ eq : vrz ] ) is a universal function of the loss in the scatterer . this work was originally stimulated by discussion with j. p. parmantier who pointed out to us the result of ref . we also acknowledge discussions with p. a. mello and d. v. savin . this work was supported in part by the dod muri for the study of microwave effects under afosr grant f496200110374 , and durip grants fa95500410295 and fa95500510240 . p. a. mello , narendra kumar , _ quantum transport in mesoscopic systems _ ( oxford university press , new york , 2004 ) . h. -j . stckmann , _ quantum chaos _ ( cambridge university press , cambridge , england , 1999 ) . f. haake , _ quantum signatures of chaos _ ( springer - verlag berlin heidelberg , 1991 ) . r. g. newton , _ scattering theory of waves and particles _ ( mcgraw - hill , new york , 1966 ) . v. pagneux and a. maurel , phys . lett . * 86 * , 1199 ( 2001 ) . e. doron , u. smilansky and a. frenkel , phys . lett . * 65 * , 3072(1990 ) . p. w. brouwer and c. w. j. beenakker , phys . b. * 55 * 4695 ( 1997 ) . u. kuhl , m. martnez - mares , r. a. mndez - snchez and h. -j . stckmann , phys . * 94 * 144101 ( 2005 ) . s. hemmady , x. zheng , e. ott , t. m. antonsen and s. m. anlage , phys . * 94 * , 014102 ( 2005 ) . y. alhassid , rev . 72 * , 895 ( 2000 ) ; c. w. j. beenakker , rev . * 69 * , 731 ( 1997 ) . b. l. altshuler et.al . _ mesoscopic phenomena in solids _ , ( north - holland , armsterdam , 1991 ) . e. p. wigner , ann . math . * 53 * , 36 ( 1951 ) ; * 62 * , 548 ( 1955 ) ; * 65 * , 203 ( 1957 ) ; * 67 * , 325 ( 1958 ) . m. l. mehta , _ random matrices _ , second edition ( academic press , 1991 ) . j. j. m. verbaarschot , h. a. weidenmller and m. r. zirnbauer phys . rep . * 129 * , 367 ( 1985 ) . c. h. lewenkopf and h. a. weidenmller , ann . of phys . * 212 * 53 , ( 1991 ) . y. v. fyodorov and h. -j . sommers , j. math . ( n.y . ) * 38 * , 1918 ( 1997 ) ; y. v. fyodorov and d. v. savin , jetp lett . , * 80 * , 725 ( 2004 ) . f. beck , c. dembowski , a. heine , and a. richter , phys . e. * 67 * , 066208 ( 2003 ) . c. fiachetti and b. michielsen , elect . 39 * , 1713 ( 2003 ) . w. hauser and h. feshbach , phys rev . * 87 * , 366 1952 . w. a. friedman and p. a. mello , ann . of physics , * 161 * , 276 ( 1985 ) . d. agassi , h. a. weidenmller , and g. mantzouranis , phys.rep . * 22 * 145 ( 1975 ) . x. zheng , t. m. antonsen and e. ott , electromagnetics , * 26 * ( 1 ) , 3 ( 2006 ) . p. a. lee and t. v. ramakrishnan , rev . phys . * 57 * , 287 ( 1985 ) ; h. u. baranger , r. a. jalabert , and a. d. stone , phys . * 70 * , 3876 ( 1993 ) ; c. w. j. benakker , rev . phys . * 69 * , 731 ( 1997 ) . a. gokirmak , d. -h . wu , j. s. a. bridgewater and s. m. anlage , rev . * 69 * , 3410 ( 1998 ) . so , s. m. anlage , e. ott and r. n. oerter , phys . rev * 74 * , 2662 ( 1995 ) . wu , j. s. a. bridgewater , a. gokimak , and s. m. anlage , phys . lett . , * 81 * , 2890 ( 1998 ) . s .- chung , a. gokirmak , d. -h . wu , j. s. bridgewater , e. ott , t. m. antonsen , and s. m. anlage , phys . lett . * 85 * , 2482 ( 2000 ) . s. hemmady , x. zheng , t. m. antonsen , e. ott , and s. m. anlage , phys . e * 71 * , 056215 ( 2005 ) .	in wave chaotic scattering , statistical fluctuations of the scattering matrix @xmath0 and the impedance matrix @xmath1 depend both on universal properties and on nonuniversal details of how the scatterer is coupled to external channels . this paper considers the impedance and scattering variance ratios , @xmath2 and @xmath3 , where @xmath4/\{var[z_{ii}]var[z_{jj } ] \}^{1/2}$ ] , @xmath5/\{var[s_{ii}]var[s_{jj } ] \}^{1/2}$ ] , and @xmath6 $ ] denotes variance . @xmath2 is shown to be a universal function of distributed losses within the scatterer . that is , @xmath2 is independent of nonuniversal coupling details . this contrasts with @xmath3 for which universality applies only in the large loss limit . explicit results are given for @xmath2 for time reversal symmetric and broken time reversal symmetric systems . experimental tests of the theory are presented using data taken from scattering measurements on a chaotic microwave cavity .
in the august 1998 issue of the _ dwarf tales _ newsletter , karatchensev and karatchenseva reported the discovery of two new probable companions to m 31 in the pegasus and cassiopeia constellations , which were identified in their visual inspection of the poss - ii film copies . one of these galaxies had been discovered independently by armandroff et al . ( as reported in the same issue ) using a digital filtering technique on poss - ii images . they provisionally designated their galaxy and vi . in the above mentioned issue of _ dwarf tales _ it was not clear which one of the two galaxies reported by karachentsev and karachentseva corresponded to and vi . in our case , this was only to be discovered right at the telescope ! at the london airport on our way to south africa , the excitement produced by the recent discovery of these new local group galaxies was clear when nobody could wait to attend the meeting to share all the new images and color - magnitude diagrams ( cmds ) . after more work and some thought , we present here our data on these new local group galaxies , together with the cmd of another local group galaxy candidate , cam a , discovered by karatchensev ( 1994 ) . peg dsph = and vi , cas dsph , and cam a were observed at the palomar 200 " telescope ( palomar observatory , usa ) on august 25 and 26 , 1998 . the seeing was excellent , 0.80.9 . only the first night , when peg dsph and cam a were observed , was photometric . a photometric transformation for cas dsph was obtained later at the 80 cm telescope iac80 at the teide observatory ( canary islands , spain ) . the photometry for all three galaxies was obtained using the daophot / allframe programs ( stetson 1994 ) . for a complete description of the observations and data reduction , see gallart et al . ( 1999b ) . peg dsph = and vi and cas dsph are highly resolved into stars in our images . their smooth , elliptical profile and their lack of blue stars and hii regions suggest that they may be two new dsph galaxies similar to the other four ( known so far ) andromeda companions , andi - iii ( caldwell et al . 1992 ) and and v ( armandroff et al . 1998 ) . their cmds , which show a well populated red - giant branch ( rgb ) but no signs of any recent star formation activity , confirm this . figure [ peg ] displays the cmd of peg dsph obtained with a total exposure time of 1 hour in both @xmath0 and @xmath1 with seeing under 0.8 . the tip of the rgb ( trgb ) appears at @xmath2 . our photometry limit is at @xmath3 , i.e. about 4 magnitudes below the trgb . therefore , the horizontal - branch ( hb ) of an old , metal - poor population is expected to be near the limit of the photometry . in fact , the widening of the lower part of the cmd may be due not only to the observational errors , but also to the fact that we are reaching the hb and red - clump of core he - burning stars , which are more extended in color than the rgb . figure [ cas ] displays the cmd of the cas dsph obtained with the same exposure time as the peg dsph . nevertheless , the fact that the night was not photometric and the seeing was slightly worse ( but still as good as 0.9 ) results in a limiting magnitude @xmath4 . the trgb is at @xmath5 . from the magnitude of the trgb , a distance similar to the distance of the andromeda galaxy ( 770 kpc , freedman & madore 1990 ) can be inferred for both peg dsph and cas dsph . this , together with the position of these galaxies in the sky , may point to a physical connection with m31 . the cmd of both galaxies indicates a predominantly old stellar population with some intermediate - age population suggested by the presence of a few stars above the trgb . in the case of the peg dsph the position and width of the rgb seem to indicate a very low metallicity ; since in that case , the agb would be less extended , we can not exclude the possibility of a relatively large intermediate - age population fraction . this is the case for the predominantly intermediate - age dsph galaxy leo i ( gallart et al . 1999a ) , whose trgb region is quite similar to the one presented by peg dsph . cas dsph appears to have a considerably larger metallicity and metallicity dispersion than peg dsph . cam a appears much less resolved in our observations than the other two galaxies , and has an irregular morphology , with a large number of blue super - giant stars and a number of apparently semi - resolved knots which may be star clusters or hii regions . its cmd in figure [ cama ] is similar to the bright part of the cmd of star forming local group galaxies like m33 , ic1613 or ngc 6822 , with a blue plume of massive main - sequence stars at @xmath6 and a red - supergiant branch at @xmath7 . the overdensity of stars fainter than @xmath8 at @xmath9 may be the red - tangle ( gallart et al . 1996 ) , observed near the limit of our photometry . a crude estimate of the distance of this galaxy can be obtained from the different features in the cmd . first , from the magnitude of the brightest stars , assuming that they would have approximately the same @xmath10 as those in m33 and ic1613 , we infer a distance modulus for cam a of @xmath11 . at this distance the trgb would be detected at @xmath4 i.e. , it would be at the limit of our photometry . alternatively , if the overdensity of stars fainter than @xmath8 at @xmath9 is indeed the red - tangle ( gallart et al . 1996 ) , it would imply a trgb at @xmath12 , and a distance modulus of @xmath13 .	we present observations of the new local group galaxy candidates cassiopeia dsph , pegasus dsph = and vi and camelopardalis a. our deep color - magnitude diagrams show that the first two galaxies are certainly local group members , and likely dsph galaxies at a distance similar to that of the andromeda galaxy . cam a seems to be a star forming galaxy situated considerably further away .
shoichi sakata was the first foreigner who visited the itep theory division . he came in the spring of 1956 and compiled a list of the itep theorists pomeranchuk , v.b . berestetsky , a.d . galanin , a.p . rudik , b.l . ioffe , v.v . sudakov , i.yu . kobzarev and myself . sakata also took a photo of those who were present . ( it would be interesting to find this picture in his archives . ) i still have the three pages of thin rice paper with the sakata model which he left with us . they correspond to his paper @xcite . these three pages were crucial for all my life in physics . sakata @xcite considered 7 mesons ( 3 @xmath0 , 4 @xmath1 ) and 8 baryons ( 2 @xmath2 , @xmath3 , 3 @xmath4 , 2 @xmath5 ) known at that time . he postulated that 3 baryons @xmath6 are more fundamental than the other 5 baryons and 7 mesons and demonstrated that these 12 particles could be composed from @xmath6 and @xmath7 . the paper had a philosophical flavor and contained no experimental predictions . in 1956 particle physicists were discussing the @xmath8-puzzle and parity violation ( see reference @xcite for further details ) . therefore the paper @xcite as well as three accompanying papers of sakata s students @xcite had no immediate response . ( s. tanaka @xcite discussed @xmath8-parity degeneracy in the sakata model , z. maki @xcite attempted to calculate bound states of baryons and antibaryons , while k. matumoto @xcite suggested a semi - empirical formula for masses of composite particles . ) in the summer of 1957 i suddenly `` reinvented '' the sakata model and realized its beauty and its potential . then i recalled the three rice pages and reread them . my first paper @xcite on the sakata model was presented by i.i . gurevich at the conference in padua venice , september 1957 . a slightly different text @xcite was published in a russian journal . in these publications the three `` sakatons '' were not physical @xmath9 , but some primary particles denoted by the same letters , so `` we can assume that for the primary particles @xmath10 '' @xcite . strong and weak interactions of sakatons were considered and for the latter a number of selection rules were deduced , in particular , those which are known as @xmath11 , @xmath12 for nonleptonic decays of strange particles via the @xmath13 transition , while for the leptonic ( or semi - leptonic ) ones @xmath11 , @xmath14 and @xmath12 via @xmath15 current . as for the strong interactions , the existence of @xmath16- and @xmath17-mesons was predicted in @xcite ; i denoted them @xmath18 and @xmath19 : in the framework of this scheme there is a possibility of two additional neutral mesons which have not so far been observed : @xmath20 the isotopic spin of the @xmath21-mesons is zero . @xcite ( the unconventional minus sign in the definition of @xmath19 was in accord with the not less unconventional definition @xmath22 . ) in december 1957 four soviet particle physicists ( d.i . blokhintsev , v.p . dzhelepov , s.ya . nikitin and myself ) visited palo alto , berkeley , boston , new york , brookhaven . for me it was my first trip abroad and the first flight in my life . ( the next time the soviet authorities allowed me to visit the usa was only in 1988 for the neutrino88 conference . ) during the 1957 trip i talked with m. baker , h. bethe , s. drell , r. feynman , r. gatto , m. gell - mann , s. goldhaber , c. sommerfield , f. zachariasen , c. zemach and many others , gave a seminar at berkeley . as a result of this seminar e. segre invited me to write an article for the annual review of nuclear science . it appeared in 1959 ( see below ) . my second paper on the sakata model `` mass reversal and compound model of elementary particles '' was published in june 1958 as a dubna preprint @xcite and i had it with me at the 1958 rochester conference at cern . on the initiative of j.r . oppenheimer and r.e . marshak a special seminar was arranged at which i presented my paper at the start of the conference and then was asked to present it also at session 7 , `` special theoretical topics '' , see @xcite . ( note that selection rules for weak interactions in sections 14 , 15 and references 24 - 28 of the dubna preprint @xcite were deleted by the editors of the proceedings @xcite ; see the appendix for the deleted pages . ) in @xcite @xmath18 and @xmath19 became mixtures of the states discussed above . what is more important , all interactions were assumed to be @xmath23-invariant following papers @xcite and especially @xcite . the conservation of the vector non - strange current , postulated in @xcite , was shown in @xcite to be inevitable in the sakata model . unfortunately the strong interaction was written as an ugly four - fermion interaction of sakatons . the discussion of my talk involved r. gatto , g. lders , r. adair , g. wentzel , t.d . lee , y. yamaguchi ( see page 228 of the proceedings ) . the discussion with yoshio yamaguchi continued during the lunch in the cern canteen . in the afternoon of the same day j. oppenheimer commented my argument that in the sakata model conservation of the weak non - strange vector current is inevitable ( see page 257 ) . he again at length commented the subject in his `` concluding remarks '' at the conference ( see page 293 ) . r. marshak stressed the novelty of chiral invariance for strong interactions ( see page 257 ) . in his talk `` @xmath24 and @xmath25 decays and related subjects '' marshak repeatedly underlined that for these decays `` @xmath26 in okun s model '' ( see @xcite , pp . 284 , 285 ) . in the discussion @xcite i described an upper limit on @xmath27 transitions which had been derived by b. pontecorvo and myself @xcite . on the basis of the selection rules for weak interactions which follow from the sakata model the lifetime of @xmath28 and its branching ratios were predicted @xcite by i.yu . kobzarev and myself . this prediction was cited by me in december 1957 at stanford and as reference [ 28 ] in the dubna preprint @xcite and was soon confirmed experimentally @xcite . in 1959 my paper @xcite appeared as well as its russian twin @xcite . i received a hundred requests for reprints , many of them from japan . strangely enough , rereading now this paper , i do not see in it the prediction of @xmath16 and @xmath17 and any statement that @xmath6 are not physical baryons , but some more fundamental particles . both the prediction and the statement were in @xcite . i can not understand now their irrational omission in @xcite . in 1959 other authors started to publish papers on the sakata model . a. gamba , r. marshak and s. okubo @xcite pointed out the symmetry between the three leptons ( @xmath29 ) and three baryons ( @xmath30 ) `` in models of sakata @xcite and okun @xcite '' . this symmetry has been emphasized by marshak ( in his rapporteur talk @xcite at the 1959 rochester conference in kiev ) and became known as the kiev symmetry . ( i served as a scientific secretary of r. marshak and participated in preparation of his report . ) at the kiev conference m. gell - mann told me : `` if i were you , i would introduce in the @xmath31 model the linear superposition ( @xmath32 ) '' . i do not understand why i did not follow his advice . the angle @xmath33 is known now as the cabibbo angle . the weak current @xmath34 first appeared next year in the paper by m. gell - mann and m. levy @xcite . in 1959 the symmetry which is now called su(3 ) was introduced into sakata model . y. yamaguchi @xcite with reference to @xcite stressed the existence of 9 pseudoscalar mesons ( @xmath35 ) . o. klein @xcite and s. ogawa @xcite discussed generalizations of isotopic symmetry . in particular , s. ogawa with reference to @xcite considered 3 doublets @xmath36 , @xmath37 , @xmath38 and 3 meson triplets . m. ikeda , s. ogawa , y. ohnuki @xcite with reference to @xcite developed some mathematical constructs of the symmetry to which they referred as u(3 ) . o. klein @xcite discussed the interaction between the triplet of sakatons and the octet of pseudoscalar mesons and stressed the symmetry between @xmath39 and @xmath40 . in 1960 i was invited to give a rapporteur talk at the rochester conference in rochester . i prepared the draft of the talk , but was not allowed by soviet authorities to attend the conference . my draft @xcite based on the sakata model has been prepared for the proceedings by s. weinberg . goldberger who `` was thrown into a breach at a rather late date '' served as a rapporteur on `` weak interactions ( theoretical ) '' @xcite referred to my draft . r. feynman @xcite spoke on the conserved vector current . he said that in the model of fermi and yang `` as has been pointed out in much more detail by okun , in any complex structure , the coupling of the beta decay is proportional to the total isotopic spin '' . m. gell - mann @xcite spoke on the conserved and partially conserved currents . he said : `` ... there is the scheme mentioned by feynman and favored by okun , marshak , and others , based on just @xmath41 , and @xmath3 . of course , if that is right we do not need the elaborate machinery i just described . we simply draw an analogy '' . but as it is clear from their talks both feynman and gell - mann at that time preferred to use the composite model only as a tool to formulate more general phenomenological approaches . among the talks at rochester 1960 was that by y. ohnuki @xcite who with a reference to @xcite assumed @xmath42 and the three - dimensional unitary symmetry . an important paper of 1960 was that by j. sakurai @xcite . with a reference to @xcite he mentioned that instead of @xmath43 one can use as `` elementary '' @xmath44 . he considered the absence of @xmath16-meson as a serious problem : `` ... within the framework of fermi yang sakata okun model it may be difficult to explain why the @xmath16 does not exist '' ( see pp . 32 - 36 ) . the @xmath16-meson was discovered within a year @xcite . further progress in su(3 ) symmetric sakata model was achieved by m. ikeda , y. miyachi , s. ogawa @xcite , who applied this symmetry to weak decays . z. maki , m. nakagava , y. ohnuki , s. sakata published a paper on sakata model @xcite . they wrote : `` ... it has recently become clear that feynman gell - mann current derived from the sakata model is quite sufficient to account for the experimental facts concerning the weak processes @xcite '' . they postulated the existence of a so - called @xmath45 matter . the bound state @xmath46 had been identified with @xmath47 , bound state @xmath48 with @xmath3 , while @xmath49 with @xmath50 . in 196061 i was giving lectures @xcite based on the sakata model . subsequently they were recast into the book @xcite . my major mistake at that time was that i did not consider seriously eight spin 1/2 baryons as an su(3 ) octet in spite of the `` eightfold way '' papers by m. gell - mann @xcite and y. neeman @xcite . ( the former referred to papers @xcite . ) in 1962 the sakata model was `` falsified '' for a short time by experiments @xcite , which discovered decays @xmath51 and @xmath52 forbidden by @xmath53 rule . at the 1962 geneva conference i tried to find a mistake in the results @xcite but failed . pomeranchuk who witnessed the argument commented later that my `` feathers were flying '' . i do not remember now how the mistake was found subsequently by experimentalists . maybe it was a statistical fluctuation . the authors of articles @xcite referred to the paper by feynman and gell - mann @xcite . while in my papers @xcite the forbidden decays were simply listed , in @xcite the notations @xmath54 and @xmath55 were used and the currents with @xmath56 and @xmath57 ( @xmath58 and @xmath59-currents ) were phenomenologically considered on the same footing . the product of these currents gives transitions with @xmath60 . the limit on these transitions from the absence of decays @xmath61 was not reliable because `` so few @xmath5 particles have been seen that this is not really conclusive''@xcite . ( the paper @xcite ( published in june 1957 ) had put a much better limit on @xmath60 processes from @xmath62 transitions . but it was not known to feynman and gell - mann when they wrote @xcite . ) in 1962 m. gell - mann predicted the existence of @xmath63-hyperon @xcite . i. kobzarev and myself @xcite derived the su(3 ) relations between semileptonic decays of @xmath0 and @xmath1-mesons . together with relations for the decays of baryons they were later derived by n. cabibbo @xcite . the discovery of @xmath64 prompted attempts to reconcile the existence of two neutrinos with the lepton - sakaton symmetry . in order to preserve the kiev symmetry z. maki , m. makagawa , s. sakata @xcite modified the @xmath45 matter model @xcite . they assumed that @xmath65 , where @xmath66 is one of the two orthogonal superpositions of @xmath67 and @xmath64 . the other superposition @xmath68 was assumed either not to form at all a bound state with @xmath45 or to form a baryon with a very large mass . on the basis of this model the paper introduced @xmath69 oscillations . another way to lepton - sakaton symmetry was suggested in the paper by y. katayama , k. motumoto , s. tanaka , e. yamada @xcite , where the fourth sakaton was explicitly introduced . in 1964 @xmath17-meson and @xmath63-hyperon were discovered @xcite . earlier this year g. zweig @xcite and m. gell - mann @xcite replaced the integer charged sakatons by fractionally charged particles ( aces zweig ; quarks gell - mann ) . this allowed them to construct not only the octet and singlet of mesons , but also the octet and decuplet of baryons . when establishing the electromagnetic and weak currents in the quark model m. gell - mann @xcite referred to similar expressions in the sakata model . on november 3 2006 i received the following email from a colleague and a friend of mine valentine i. zakharov : dear lev borisovich , i am now visiting kanazawa , japan . this month , there will be a one - day conference in nagoya , to celebrate 50 years of the sakata model . they invited me to come and i eagerly agreed . one of the reasons which you can readily guess was that the words ` sakata model ' were among the first ones i heard about our field . ( you were giving lectures to ` experimentalists ' , with alikhanov in the first row ; ( m.i . ryazanov from mephi encouraged me to attend ; it was some time before i showed up later ) . i will mention of course that you were developing the sakata model at itep . but , unfortunately , i realized that i do not know anything else , to any extent personal about sakata - sensei . i mean , no other papers , or their echo in russia / ussr , nothing ... may be you can help in some way ? excuse me , please , for bothering you and with best regards , valya . to answer valentine s request i have written this brief review . thinking that it might be of more general interest , i published it as version 1 of hep - ph/0611298 . on december 22 i received an email from k. yamawaki who kindly invited me to publish this paper in the proceedings of the sakata model symposium . in editing the paper i benefited from email exchanges with s. pakvasa , h. lipkin and a. gal . another interpretation of the terms `` sakata model '' and `` sakata symmetry '' one can find in preprints @xcite . four pages from the dubna preprint @xcite : 99 s. sakata . on a composite model for new particles . . phys . * 16 * ( 1956 ) 686 . mirror particles and mirror matter : 50 years of speculation and search . hep - ph/0606202 , v.2 . k. matumoto . some consequences of the compound hypothesis of elementary particles . * 16 * ( 1956 ) 583 . s. tanaka . the composite model for new unstable particles , i. progr . theor . phys . * 16 * ( 1956 ) 625 . z. maki . on a theory of a composite model of elementary particles . * 16 * ( 1956 ) 667 . some remarks concerning the compound model of fundamental particles . proc . of the intern . conf . on mesons and recently discovered particles , padova - venezia , 22 - 28 settembre 1957 , p. v-55 . some remarks on a compound model of elementary particles . jetp * 7 * ( 1958 ) 322 ; zhetf * 34 * ( 1958 ) 469 ( in russian ) . mass reversal and compound model of elementary particles . dubna preprint p-203 ( 1958 ) . mass reversal and compound model of elementary particles . proc . of 1958 annual international conference on high energy physics at cern , 30th june-5th july , 1958 , p. 223 . feynman , m. gell - mann . theory of the fermi interaction . phys . rev . * 109 * ( 1958 ) 193 . sudarshan , r.e . chirality invariance and the universal fermi interaction . phys . rev . * 109 * ( 1958 ) 1860 . sudarshan , r.e . the nature of the four fermion interaction . proc . of the intern . conf . on mesons and recently discovered particles . padova - venezia , 22 - 28 settembre 1957 , p. v-14 . mass reversal and weak interactions . nuovo cimento * 7 * ( 1958 ) 649 . gershtein , ya.b . zeldovich . on meson corrections in the theory of @xmath70-decay . jetp * 2 * ( 1956 ) 576 ; zhetf * 29 * ( 1955 ) 698 ( in russian ) . r.e . marshak . @xmath71 and @xmath72 decays and related subjects . proc . of 1958 annual international conference on high energy physics at cern , 30th june-5th july , 1958 , p. 284 . of 1958 annual international conference on high energy physics at cern , 30th june-5th july , 1958 , p. 201 . l. okun , b. pontecorvo . some remarks on slow processes of transition of elementary particles . jetp * 5 * ( 1957 ) 1297 ; zhetf * 32 * ( 1957 ) 1587 ( in russian ) . kobzarev , l.b . lifetime of the @xmath73 meson . jetp * 7 * ( 1958 ) 524 ; zhetf * 34 * ( 1958 ) 763 ( in russian ) . m. bardon , m. fuchs , k. lande , l.m . lederman , w. chinowsky , j. tinlot . lifetime and decay of the @xmath28 meson . phys . rev . * 110 * ( 1958 ) 780 . strange particles : decays . annual review of nuclear science * 9 * ( 1959 ) 61 l. okun . strange particles : decays . uspekhi fiz . nauk * 68 * ( 1959 ) 449 ( in russian ) . a. gamba , r. marshak and s. okubo . on a symmetry in weak interactions . * 45 * ( 1959 ) 881 . . theoretical status of weak interactions . ninth international annual conference on high energy physics , v.2 , moscow , 1960 , p. 269 . m. gell - mann , m. levy . the axial vector current in beta decay . nuovo cimento * 16 * ( 1960 ) 705 . . a composite theory of elementary particles . supplement to the progr . * 11 * ( 1959 ) 1 . o. klein . on the systematics of elementary particles . archiv fr fysik swed . * 16 * ( 1959 ) 191 . s. ogawa . a possible symmetry in sakata s composite model . * 21 * ( 1959 ) 209 . m. ikeda , s. ogawa , y. ohnuki . a possible symmetry in sakata s model for bosons - baryons system . * 22 * ( 1959 ) 715 certain problems of weak interaction theory . proceedings of rochester conference , august 25 - sept . 1 , 1960 , p. 743 . weak interactions ( theoretical ) . proceedings of rochester conference , august 25 - sept . 1 , 1960 , the status of the conserved vector current hypothesis . proceedings of rochester conference , august 25 - sept . 1 , 1960 , p , 502 . m. gell - mann . conserved and partially conserved currents in the theory of weak interactions . proceedings of rochester conference , august 25 - sept . 1 , 1960 , y. ohnuki . composite model of elementary particles . proceedings of rochester conference , august 25 - sept . 1 , 1960 , sakurai , theory of strong interactions . annals of physics * 11 * ( 1960 ) 1 . a. pevsner et al . evidence for three pion resonance near 550 mev . * 7 * ( 1961 ) 421 . l.b . okun . lectures on the theory of weak interactions of elementary particles . 17 itep preprints , 1960 - 1961 ( in russian ) . english translation : theory of weak interactions : thirteen lectures , aec - tr-5226 . us atomic energy commission . oak ridge , tenn . ( for the translation of lectures 14 - 16 and of contents see np-10254 , 10842 , 10845 , 10840 . ) lectures on the theory of weak interactions of elementary particles . jinr preprint p-833 ( 1961 ) ( in russian ) . slaboe vzaimodeistvie elementarnykh chastits . moskva , fizmatgiz , 1963 ( in russian ) . english translation : weak interaction of elementary particles . pergamon press , 1965 . m. ikeda , y. miyachi , s. ogawa . symmetry in sakata s model and weak interactions . * 24 * ( 1960 ) 569 . z. maki , m. nakagava , y. ohnuki , s. sakata . a unified model for elementary particles . * 23 * ( 1960 ) 1174 . s.okubo , r.e . marshak , e.c.g . sudarshan . v - a theory and decay of @xmath3 hyperon . phys . rev . * 113 * ( 1959 ) 944 . m. gell - mann . the eightfold way : a theory of strong interaction symmetry . report ctsl-20 ( 1961 ) . y. neeman . derivation of strong interactions from a gauge invariance . * 26 * ( 1961 ) 222 . r. ely et al . experimental test of the selection rule @xmath53 . * 8 * ( 1962 ) 132 . g. alexander et al . experimental test of the @xmath74 rule , and @xmath75 rule in three - body decays of neutral @xmath1 mesons . * 9 * ( 1962 ) 69 . m. gell - mann . strange particle physics . strong interactions . proc . intern . . high energy phys . ( cern , 1962 ) , p. 805 . m. gell - mann . symmetries of baryons and mesons . * 125 * ( 1962 ) 1067 . kobzarev , l.b . unitary symmetry and universal weak interaction . jetp * 15 * ( 1962 ) 970 ; zhetf * 42 * ( 1962 ) 1400 ( in russian ) . n. cabibbo . unitary symmetry and leptonic decays . phys . * 10 * ( 1963 ) 531 . z. maki , m. nakagava , s. sakata . remarks on the unified model of elementary particles . * 28 * ( 1962 ) 870 . y. katayama , k. matumoto , s. tanaka , e. yamada . possible unified models of elementary particles with two neutrinos . * 28 * ( 1962 ) 675 . g. danby , j.m . gaillard , k. goulianos , l.m . lederman , n.b . mistry , m. schwartz , j. steinberger . high energy neutrino interaction in matter . proc . of ichep 1962 at cern , p.m. dauber et al . properties of the 960-mev boson . * 13 * ( 1964 ) 449 . barnes et al . observation of a hyperon with strangeness minus three . * 12 * ( 1964 ) 204 . an su(3 ) model for strong interaction symmetry and its breaking . preprints cern - th-401 , 412 . m. gell - mann . a schematic model of baryons and mesons . * 8 * ( 1964 ) 214 . h.j . lipkin . from sakata model to goldberg neeman quarks and nambu qcd phenomenology and `` right '' and `` wrong '' experiments . hep - ph/0701032 and a revised version to appear in this proceedings . the hypernuclear physics heritage of dick dalitz ( 19252006 ) . hep - ph/0701019 .	the evolution of the sakata model is described on the basis of personal recollections , proceedings of international conferences on high energy physics and some journal articles .
the discovery that galactic bulges harbour supermassive black holes whose masses are correlated with the properties of the bulge suggests a close connection between the formation of galaxies and the formation of black holes . the mass of the galactocentric black hole varies approximately linearly with the stellar mass of the bulge ( @xcite ; @xcite ; @xcite ) and roughly as the fourth power of the bulge velocity dispersion ( @xcite , @xcite , @xcite ) . correlations with the near infrared bulge luminosity ( @xcite ) and with the bulge light concentration ( @xcite ) have also been found . these correlations are remarkable because they link phenomena on widely different scales - from the parsec scale of the black hole s sphere of influence to the kiloparsec scale of bulges - and thus point to a connection between the physics of bulge formation and the physics of black hole accretion and growth ( @xcite ) . the simplest interpretation is that both , black hole and bulge growth , are driven by the same process whose nature , however , remains unclear . various models for the growth of black holes in galaxies have been studied ( e.g. @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite , and others ) . in the context of a hierarchical cold dark matter universe , a plausible explanation for the tight correlation between bulge and black hole properties is that the galaxy mergers or disc instabilities that induce bulge growth via bursts of star formation also feed the central black hole ( @xcite ; @xcite ; @xcite . ) a simple implementation of this model follows from assuming that , as cold gas condenses into stars , a certain percentage of the gas is forced into the centre of the galaxy and accreted by the black hole . models based on this and related prescriptions successfully reproduce the relation ( hereafter we use this term to refer to the relation between black hole mass and stellar bulge mass ) ( @xcite ; @xcite ; @xcite ) . an interesting aspect of the correlations between black hole mass and bulge properties is that they seem to apply over a range of five to six orders of magnitude in black hole mass ( @xcite ; @xcite ) . if a simple model of the kind just mentioned for the simultaneous growth of black holes and bulges is correct , then there are two direct consequences which we explore in this paper . the first is that black holes should exist in bulges of all luminosities including dwarf ellipticals and satellites of brighter galaxies like the milky way . the second is that black holes will likely merge as their hosts bulges collide . binary black holes orbiting each other emit gravitational waves ( @xcite ) . using quasi - newtonian methods to study the orbital decay due to gravitational wave emission , @xcite found that the system will eventually enter a plunge phase , causing the black holes to coalesce emitting a burst of gravitational waves . @xcite found that , in addition to transferring energy out of the emitting system , gravitational radiation can also take with it linear momentum . as a result , the centre of mass of the system recoils in a direction specified by the boundary conditions of the last stable orbit . the astrophysical implication of this linear momentum kick ( the `` rocket effect '' or `` gravitational recoil '' ) is that black holes may be ejected from galactic bulges if the potential is shallow and the kick is large enough ( @xcite ; @xcite ; @xcite ) . in theory , this could lead to a sizable population of extragalactic black holes which could , in principle , dominate the black hole mass function . our aim in this work is to examine the importance of such kicks for the galactic black hole population . in particular , we consider the scatter on the relation in an attempt to constrain the relatively uncertain kick velocity , as well as the nature and spatial distribution of a possible extragalactic population of ejected black holes . we model the growth of black holes using the semi - analytic galaxy formation model of @xcite , using two methods for obtaining merger trees : monte carlo techniques and high resolution n - body simulations ( in which the trajectories of recoiling black holes can be tracked . ) this paper is organised as follows . in section [ gravitaionalrecoil ] , we review the physics of the gravitational recoil . in section [ gfintro ] , we describe how we model the formation and ejection of black holes in our semi - analytic model . in section [ standalone ] , we determine the effect of black holes ejected from the progenitors of present day galaxies on the relation . in section [ nbodyintro ] , we use a set of high resolution n - body simulations of galactic halos to track the location of ejected black holes . we conclude in section [ conclusion ] where we discuss the possible consequences of an extragalactic black hole population . the emission of gravitational radiation is a generic feature of any massive asymmetrically collapsing system ( @xcite ) . in order to be accompanied by a linear momentum kick , the gravitational radiation must be asymmetric ( @xcite ) . for two coalescing black holes of unequal mass this occurs as the gravitational radiation from the lighter , more rapidly moving partner is more strongly beamed . as two black holes orbit each other , gravitational radiation causes their orbits to shrink and circularise until the last stable circular orbit is reached . using perturbation theory in the weak field approximation , @xcite calculated the kick velocity from this final orbit to be : @xmath2 where @xmath3 is the radius of the innermost stable circular orbit , @xmath4 is the mass ratio with the convention @xmath5 , and @xmath6 which reaches a maximum , @xmath7 , for a mass ratio of @xmath8 . unfortunately , the weak field approximation used by fitchett becomes invalid as the binary approaches the plunge phase and it is here that the contribution to the recoil velocity becomes largest . two decades later , favata , hughes & holz ( 2004 ) used perturbation theory to show that , in the presence of strong fields , gravitational kicks are not expected to exceed @xmath9 . in a companion paper , @xcite included the effect of the larger partner s spin and obtained an upper limit of @xmath10 . more recently blanchet , qusailah & will ( 2005 ) performed a higher order calculation of the recoil for the special case of the coalescence of two nonrotating black holes , obtaining an upper limit of @xmath11 . however , none these calculations apply to all mass and spin ratios and kick velocities as large as @xmath12 can not be definitively ruled out . in light of the uncertainty in the recoil velocity , we simply assume that the kick is directly proportional to fitchett s scaling function @xmath13 and write the recoil velocity in terms of a prefactor velocity @xmath14 , @xmath15 we then allow @xmath14 to vary between @xmath16 ( i.e. no kick ) to @xmath17 with the aim of constraining this parameter empirically . in this section , we briefly explain how we use the semi - analytic galaxy formation model ` galform ` described in detail by @xcite , @xcite and @xcite to model the growth and evolution of galaxies . the growth of dark matter halos by mergers is encoded in a merger tree . along each branch of the tree , the following physical processes are calculated : ( i ) shock - heating and virialization of gas within the gravitational potential well of each halo ; ( ii ) radiative cooling of gas onto a galactic disc ; ( iii ) the formation of stars from the cooled gas ; ( iv ) the effects of photoionization on the thermal state and cooling properties of the intergalactic medium ; ( v ) reheating and expulsion of cooled gas through feedback processes associated with stellar winds and supernovae explosions ( see @xcite ) ; ( vi ) the evolution of the stellar populations ; ( vii ) the effects of dust absorption and radiation ; ( viii ) the chemical evolution of the stars and gas ; ( ix ) galaxy mergers ( which , depending on the violence of the merger , may be accompanied by starbursts and the formation of a bulge see @xcite ) ; ( x ) the evolution of the size of the disc and bulge . we used two different methods for obtaining dark matter halo merger trees . in the first instance , we constructed the trees using a monte - carlo algorithm to determine merger rates ( e.g. @xcite , @xcite ) . in the second instance , we explicitly extracted the merger trees by following the merging of dark matter halos in n - body simulations ( @xcite ) . the latter method allows us to apply the semi - analytic formalism directly to the simulations . more details are given in section [ demographics ] . stellar discs are formed as rotating diffuse halo gas cools and settles in a halo . in the ` galform ` model , discs are assumed to have an exponential surface density profile , @xmath18 where @xmath19 is a characteristic disc length and @xmath20 is the central surface density . the potential of such a mass distribution as a function of just radius ( i.e. in the plane of the disc ) is @xmath21\ ] ] where @xmath22 . the functions @xmath23 and @xmath24 are modified bessel functions of the first and second kind ( e.g. @xcite ) . differencing the potential at @xmath25 and @xmath26 , we find that the escape velocity from the centre of the disc is @xmath27 where @xmath28 is the half mass radius of the disc . in the ` galform ` model , the principal mechanism for forming both elliptical galaxies and the bulges of luminous spiral galaxies is galaxy mergers . when two galaxies of comparable mass merge , this is termed a major merger and all the stars from both merging partners are assumed to form a single spheroid while any gas present is consumed in a burst of star formation . in minor mergers there is no burst of star formation ; only the stars from the smaller galaxy are added to the bulge of the larger galaxy , while the gas is added to the disc of the larger galaxy . in the context of the ` galform ` model , @xcite also considered the possibility that dynamical instability in selfgravitating discs could result in the formation of galactic bulges ( see also @xcite ) . they found that this mechanism was unlikely to contribute significantly to the formation of large bulges , but could be important in creating the bulges of lower luminosity galaxies . the criterion they used to determine whether a cold disc was unstable was @xmath29 , where @xmath30 @xcite . here @xmath31 is the circular velocity at the disc half - mass radius , @xmath28 . we adopt the slightly lower value of @xmath32 in order to obtain a distribution of bulge to total luminosities similar to that found in the sloan digital sky survey by @xcite . for such unstable discs it was assumed that all the disc stars would be transformed into bulge stars and any cold disc gas would undergo a burst of star formation . the assumption that bulges , and as we discuss later , black holes grow as a result of discs becoming unstable is also assumed in the agn feedback models incorporated in the latest semi - analytic models of @xcite and @xcite . to illustrate the importance of this mechanism , we present models both with and without this secondary route for generating galactic bulges . the density profile we adopt for bulges is the @xcite model , @xmath33 where @xmath34 is a scale length and @xmath35 is the total bulge mass . the potential of such a profile is found by integrating the poisson equation and the escape velocity from the centre of such a spheroid is given by @xmath36 where @xmath37 is the bulge s half mass radius . we assume that the growth of black holes is proportional to the growth of bulges . whenever an event that adds stellar mass to the bulge occurs , we assume that a fraction , @xmath38 , of that mass is accreted by the central black hole . this assumption is motivated in part by models of black hole growth such as that of @xcite , wherein black holes accrete a proportion of the cold gas being consumed in a burst of star formation . in our model , black holes grow both as a result of such star formation bursts _ and _ also by the accretion of stars in mergers or as the result of disc instabilities . we assume that in each case the same mass fraction , @xmath38 , is channelled onto the black hole . this simplifying assumption has the virtue that , in the absence of black hole ejection , all bulges will have black holes that sit precisely on the relation and so we can , in principle , use the scatter induced by black hole ejections to set a firm upper limit on the kick velocities . the precise value of @xmath38 needed to match the data has changed in the literature ( e.g. see @xcite ; @xcite ) and we choose @xmath39 , the value published by @xcite from estimates of black hole masses for 72 active galaxies . during a merger , we assume that a fraction of any new stellar bulge mass is added to the existing black hole , @xmath40 if the black hole is ejected during the merger , we assume that a new black hole is born whose mass is equal to the mass fraction that would have been added to the preexisting black hole had it not been ejected , that is : @xmath41 this black hole then becomes the focus of mass accretion during subsequent episodes of bulge growth . in this way , we ensure that all bulges contain a black hole at all times and that , in the absence of gravitational recoil , all black holes at all times have masses directly proportional to their bulges and hence lie on a perfect relation with zero scatter . the implicit assumption that black holes exist only in bulges and not in discs , and that black holes grow during mergers rather than during quiescent star formation is motivated simply by the desire to reproduce the empirical relation . we assume that once two galaxies have merged , their central black holes coalesce instantaneously . this should be a good approximation as the black holes will merge on a dynamical timescale while the time between galaxy mergers , which is determined by the hierarchical growth of structure , is typically much longer . note that even when the merger rate of dark matter halos is relatively high , the galaxies orbiting inside the dark halos will only merge once their orbits decay due to energy loss by dynamical friction against the halo material ( @xcite ) . this implies that three - body mergers in which ejections could occur via a slingshot effect ( e.g. @xcite ) are expected to be rare . the mass ratio of the merging black holes determines the recoil velocity , @xmath42 , according to equation ( [ eqvkick ] ) . thus , to determine whether the recoiling black hole escapes from the galaxy we add the bulge and disc potentials and require @xmath43 note that we neglect the contribution of the dark matter halo . this is a good approximation over the lengthscale of the galaxy because the dark matter potential has a very shallow gradient . however , the dark matter potential becomes increasingly important for black holes that escape the galaxy and , in section [ demographics ] , we use an alternative method to follow their orbits . if the recoil velocity , @xmath42 , is less then @xmath44 we assume that the black hole will not escape but rapidly return to the centre of the galaxy . this should be a good approximation : according to the calculations of @xcite , recoiling black holes that have insufficient kinetic energy to escape undergo damped oscillations about the galactic nucleus which decay in a few dynamical times . our modelling of black hole growth thus automatically reproduces an empirical relation . if the prefactor velocity , @xmath14 , in equation ( [ eqvkick ] ) is zero , the black hole masses are all directly proportional to the bulge mass of their host galaxies . if @xmath14 is not zero , gravitational kicks introduce scatter in the relation which we now investigate and compare with observations . in order to investigate how the relation is affected by the ejection of recoiling black hole merger remnants we have studied a sample of 1000 halos of final mass @xmath45 . merger trees for each halo were generated using the monte - carlo method described by @xcite and the ` galform ` rules of galaxy formation were applied to each branch of the tree . the growth of each central black hole ( which tracks the growth of the bulge ) was calculated as described in section 3.3 . our fiducial model includes the process of bulge formation by disc instability discussed in section 3.2 . for every black hole merger we consider kick velocities corresponding to values of the prefactor , @xmath14 , in eqn ( 2 ) in the range @xmath46 - @xmath17 . relation for the case in which black hole merger remnants have the kick velocity given by equation ( [ eqvkick ] ) with @xmath47 . the diagonal line represents the ` ideal ' relation for which @xmath48 . ] the ejection of a black hole from the bulge is a source of scatter in the relation . [ magplot ] shows the @xmath49 relation for @xmath50 halos assuming a moderate prefactor kick velocity of @xmath51 . note that black hole mass may only be scattered downwards by ejections . galactic bulges that have never experienced an ejection lie exactly on the diagonal line . in the absence of kicks , the model , by definition , gives a perfect relation with no scatter . as the kick velocity increases , a tail of bulges that host black holes of reduced mass appears , while the remaining galaxies whose progenitors never lost their black hole still lie precisely on the relation . the form of the scatter about the relation and its dependence on the prefactor @xmath14 are shown in fig . [ residplot ] . the upper two panels show how deviations from the ideal relation depend on bulge mass for both @xmath47 and @xmath52 , while the lower two panels show histograms of these deviations averaged over all bulge masses . we see in these lower panels that the distribution of deviations from the ideal relation is very non - gaussian , with a tail extending to very low values of log@xmath53(@xmath54 ) . for @xmath55 , 2.6% of bulges have black holes with mass smaller than 20% of the ideal value . for @xmath56 this fraction jumps to 18.9% . the rms width of this distribution is not a good statistical description of the scatter in the relation since the rms is dominated by the noisy tail . we use instead a more robust measure of width . we calculate the widths that contain @xmath57% and @xmath58% of the distribution , and then define @xmath59 and @xmath60 to be half and one third of these widths respectively . for a gaussian distribution both of these measures would equal @xmath61 , the usual rms width . [ scat ] shows how these measures of scatter vary as a function of @xmath14 and compares them to published observational determinations . the non - gaussian nature of the scatter results in the two curves being significantly different . we immediately see that the contribution to the scatter in the relation from ejected black holes is very sensitive to assumptions concerning disc stability . for the case where unstable discs are assumed to form bulges , large values of @xmath14 are strongly ruled out as they would produce more scatter in the relation than is observed . [ scat ] shows that to be consistent with the tightest observational limits ( @xcite ; @xcite ; @xcite ) requires @xmath62@xmath63 . this constraint is consistent with the kick velocities predicted by the general relativistic analysis of @xcite and @xcite . even so , if the prefactor velocity lies in the range @xmath64 - @xmath65 , as these calculations suggest , then gravitational recoil provides a substantial contribution to the scatter in the relation . in models where discs are assumed to be stable to the formation of a bulge , all the observational estimates of the scatter lie well above the values that result from recoiling black hole ejections . in this case , we conclude that even kicks with velocities exceeding the range expected from current calculations do not produce a significant contribution to the scatter in the relation . there are two reasons for such a large difference between the two models . firstly , when discs are assumed to be stable there are far fewer ( @xmath66 ) bulge - bulge mergers and far more disc - bulge and disc - disc mergers involved in the formation of the galaxy . while the total number of mergers is the same in both models , the effect of the gravitational recoil depends only on the number of bulge - bulge mergers because , in our model , black holes only grow if there is a stellar spheroid . the second reason for the difference between the two cases is that when discs are assumed to be stable , most ejections occur at early times ( @xmath67 ) when the bulges were smaller . the bulge then has enough time to regrow a black hole of the appropriate mass by subsequent galaxy mergers from which black hole ejection becomes increasingly difficult . when unstable discs are assumed to form bulges , the majority of the ejections occur much later ( @xmath68 ) and subsequent mergers are unable to grow a large black hole . note that no black holes are ejected at @xmath69 and @xmath70 for disc - stable and disc - unstable models respectively . . the curves show our estimates of the width of the distribution of @xmath54 induced by black hole ejections as discussed in the text . the solid and dashed curves represent @xmath59 and @xmath60 respectively . the thick lines show the estimates when disc instability produces bulges , while the thin lines represent models when this route for bulge formation is ignored . the horizontal lines show estimates of the scatter from various observational studies . note that tremaine et al . ( 2002 ) give only an upper limit , while marconi & hunt ( 2003 ) present two estimates to which they refer as conservative and unconservative . ] in this section , we use high resolution n - body simulations of galactic halos to track the location of subgalactic and galactic black holes as the halo grows by mergers . the simulations provide spatial information that is not available in monte - carlo merger trees . we first describe the simulations and the way in which we model black hole escapees . we have used six @xmath71-body simulations of galactic - size dark matter halos of final mass @xmath72 to study the demographics and spatial distribution of their black hole population , including black hole ejections . the simulations were performed using the ` gadget ` code ( @xcite ) in a flat @xmath73cdm universe with @xmath74 . these simulations have been previously studied by other authors ( @xcite ; @xcite ; @xcite ; @xcite ) and we refer the reader to these papers for technical details . briefly , each simulation follows the formation of structure in a cube of comoving side 35.325 @xmath75mpc with a lagrangian ` high resolution' region around the halo of interest in which the particle mass is @xmath76 and ` low resolution ' particles elsewhere . table [ tab ] summarises the parameters of our six simulations . .parameters for the six @xmath71-body halo simulations . the columns give : ( 1 ) halo label ; ( 2 ) total number of particles in the simulation cube ; ( 3 ) number of high resolution particles ; ( 4 ) virial radius of the halo in @xmath75 mpc defined as the distance from the centre to the radius at which the mean interior density is @xmath77 times the critical density ; ( 5 ) number of particles within the virial radius . all halos were simulated in a cube of comoving length @xmath78mpc in a @xmath73cdm universe , with a particle mass of @xmath79 in the ` high resolution ' region . [ cols= " < , < , < , < , < , < " , ] [ table : windows ] black holes that remain within their host bulge are always assumed to be located at the centre of the galaxy . when a black hole is ejected from its galaxy as a result of a recoil , we search for a particle in the simulation in the neighbourhood of the galaxy with appropriate velocity which , we assume , tracks the orbit of the black hole . to find as close a match as possible , we define a cost function and attach the escaping black hole to the particle that minimises the cost . the cost function we adopt is : @xmath80 where @xmath81 and @xmath82 are the position and radial velocity of the @xmath83th particle relative to the initial black hole position and velocity , @xmath84 is the velocity the black hole has when it escapes from the galaxy and @xmath85 mpc . for modest kick velocities , @xmath86 , a suitable particle is usually found and 75% of the ejected black holes are identified with particles with @xmath87 . as @xmath14 is increased , the identification becomes increasingly difficult . black holes with such large kick velocities , however , are likely to be ejected from the halo in any case . thus , if no particle with @xmath88 is found , we assume that the black hole has been lost from the halo to become part of a population of intergalactic black holes . the general properties ( number density , luminosity function ) and spatial distribution of the satellite population depend somewhat on whether or not we assume that unstable discs generate galactic bulges . the statistics of the associated black holes , on the other hand , depend strongly on this assumption since many more bulges and black holes are formed when disc instability is taken into account . as above , in what follows , we take the unstable disc case as our fiducial model , but summarise also results when only mergers are assumed to give rise to bulges . the spatial distribution of black holes and dark matter in one of our simulations is illustrated in fig . [ dotplot ] . the top two panels correspond to the model with unstable discs and the lower two to the model without unstable discs . in both cases , the left - hand column shows results for a kick velocity prefactor of @xmath47 and the right - hand column for a kick velocity prefactor of @xmath89 . the greater efficiency of bulge formation in the unstable disc case is reflected in the larger number of black holes in this model . those black holes that are still associated with a satellite galaxy are indicated by circles while those that have been ejected ( the `` wandering '' black holes ) are indicate by squares . tests of the mass resolution of our calculation in which we artificially increased the minimum halo mass used in the merger trees indicate that our catalogues are essentially complete for satellite galaxies with v - band luminosity brighter than -7 ( corresponding to a mass larger than @xmath90 ) . to this limit we find that , for the unstable disc model , each simulated galaxy halo at @xmath49 contains , on average , 52 satellites within its virial radius . the median bulge stellar mass of this population is @xmath91 with a corresponding black hole mass of @xmath92 in the ideal relation . the fraction of these satellites that retain a central black hole at the final time depends weakly on the assumed prefactor for the kick velocities . for @xmath93 85.4% of satellites retain a central black hole at the final time , but this fraction is reduced to 84.2% for @xmath94 . the number density profiles of the black hole populations of each of our simulations , assuming our fiducial @xmath95 , are shown by the dotted lines in fig . [ densprof ] . these can be compared with the dark matter density profiles and the corresponding number density profiles of satellite galaxies with a v band magnitude @xmath96 . in order to compare wandering and satellite black hole populations of similar masses , for this plot we selected only black holes with mass greater than that of the black hole in the lowest mass bulge of the satellite sample . the density profile of black holes is seen to be intermediate between that of the dark matter and that of the satellites and , in a couple of cases , wandering black holes are found as far in as @xmath97 . our understanding of this behaviour is that these black holes are ejected from the progenitors of both present day satellite galaxies and the progenitors of galaxies which by the present have merged with the central galaxy . as the ejection velocities are not large , these wandering black holes initially have orbits similar to the galaxies that ejected them but , being lighter , they are not subject to dynamical friction and so , unlike the satellite galaxies , those near the centre are not removed by merging with the central galaxy . while the right hand panel shows the same quantity for @xmath98 . the upper two panels show histograms for models where discs are unstable to bulge formation while the bottom two panels show data for models where instability is ignored . the red dashed histogram represents the mass function of black holes that reside in satellite galaxies ; the blue solid histrogram represents the mass function of wandering black holes that have been ejected from their host bulges . the black histogram is the combined mass function . ] , the right two are for @xmath99 . the upper two panels are for unstable disc models , while the lower two panels are for stable disc models . ] in the case of the fiducial model ( in which unstable discs form bulges ) , each milky - way mass halo generated on average 80 black holes of mass greater than @xmath100 for @xmath101 during the course of its formation . of these , at @xmath49 approximately half were wandering black holes and the other half were galactic black holes retained by their host satellite galaxies . of the wandering black holes , the fraction that are ejected with sufficient velocity to completely escape the halo depends more sensitively on @xmath14 . for @xmath102 , 42% become intergalactic wandering black holes while for @xmath103 this fraction increases to 60% . these numbers and the corresponding ones for the model in which disc instability is ignored are given in table 2 . at the present day , there are only 50 black holes more massive than @xmath100 within the central 100 @xmath75kpc of the halo . black hole passages through the disc are therefore rare and unlikely to have affected the structure of the disc substantially . a large population of intergalactic black holes exists beyond the virial radius of the halo . some of these are attached to small galaxies that will eventually become satellites , but others are black holes that were ejected from the halo altogether . ( only the fraction of these that were tagged with a particle are shown in fig . [ dotplot ] . ) fig . [ totalmf ] shows the mass function of halo black holes , averaged over all six simulations , for two assumed values of @xmath14 in the models with and without unstable discs . black holes in satellites within the virial radius whose host galaxy is brighter then @xmath96 and all wandering black holes are included . in all cases , the central supermassive black hole of mass @xmath104 has been omitted from the sample . in the unstable disc case , the medians of the distribution are @xmath105 and @xmath106 for @xmath107 and @xmath52 respectively . in both cases , we expect a typical halo to have @xmath108 black holes of mass @xmath109 . the mass distribution is slightly wider for the larger value of @xmath14 , reflecting the larger efficiency of black hole ejection in this case . in the case where disc instabilities are ignored , their are fewer black holes ( note the difference in the scales of the y - axes in the upper and lower panels of this figure ) , but the distributions are shifted to slightly higher black hole masses with a median of @xmath110 and virtually no black holes of mass @xmath111 . the reason why the median is shifted to slightly higher masses in the stable disc case is because in this case there are fewer ejections , thus allowing satellite galaxy black holes to grow larger . also shown in fig . [ totalmf ] is the result of splitting the black hole mass function into the two populations : the satellite galaxy black holes and wandering black holes ( red dashed and blue solid line respectively ) . the bimodal nature of the total mass function is visible as the result of these two distinct populations . the median of the satellite black hole mass distribution is @xmath112 while the wandering black hole population peaks at @xmath113 for unstable disc modes . in the stable disc models , the bimodality is also visible although slightly less pronounced . this tendency for the wandering black holes to be more massive than those retained in the satellites is shown clearly in fig . [ fracinsats ] , which plots the fraction of black holes of each mass that are associated with satellite galaxies as a function of black hole mass . we see that below @xmath114 the fraction associated with satellite galaxies is a decreasing function of black hole mass . for low masses ( @xmath115 ) , more than 90% of black holes are in satellites . however , at higher masses ( @xmath116 ) , the majority of black holes are found to be wandering . the reason for this bias in the masses of wandering black holes compared to satellite black holes is the fact that larger satellites are more likely to merge and experience a kick then smaller ones . since the dynamical friction timescale is inversely dependent on satellite mass ( eq . 4.16 in @xcite ) , we expect the physics of kicks to have a larger effect on the higher mass satellites and bulges than on the lower mass ones . the lower mass satellites will have experienced few if any ejections and their central black holes will be unaffected by velocity kicks . in principle , a large population of wandering black holes could affect the match between the total mass in black holes at the present day and the mass in quasar remnants inferred from energy considerations ( @xcite ) . in our calculations , however , wandering black holes make up , on average , only 2.6% and 3.9% of the total black hole mass in our simulated galactic halos for @xmath102 and @xmath117 respectively . wandering black holes do , however , make up a large fraction of the intrahalo black hole population ( ie . all halo black holes excluding the central one ) . for these two values of @xmath14 , unattached wandering intrahalo black holes make up 31.8% and 48.4% of the total intrahalo satellite black hole mass . another noteworthy feature of the black hole mass functions of fig . [ totalmf ] is the small population of very massive black holes with @xmath118 . this population is primarily composed of wandering black holes , although a portion of it in the unstable disc model consists of black holes in massive satellites whose bulge masses are @xmath119 , as evidenced by the high mass peak in the upper panels of fig . [ fracinsats ] . in stable disc models , this population is composed entirely of wandering black holes . these massive wandering black holes originate from ejections from the central galaxy that occured when the central galaxies bulge had a mass of @xmath120 . the central galaxy has since undergone subsequent phases of bulge formation to bring its mass to @xmath121 and , in the process , has regrown a central black hole of mass @xmath122 . the gravitational recoil of merging black holes is an important physical effect in a universe built up hierachically through the repeated merging of galactic units . whenever galaxies that host black holes merge , the black holes themselves will coalesce and there exists the potential for the remnant black hole to be ejected from the galaxy , provided the recoil velocity is high enough and the galactic potential shallow enough . we have examined the role that gravitational recoil plays on the demographics of black holes . we find that the process of ejecting black holes from galaxies is efficient if bulges and black holes grow both in galaxy mergers and as a result of discs becoming unstable , because in this case nearly all bulges contain a black hole . in models where disc instability is ignored , this process is not as efficient because fewer bulges and associated black holes exist and because black hole - black hole mergers tend to occur late when the galactic potential wells are deeper . in the former case , black hole ejections produce a significant contribution to the scatter in the relation . in fact , conservative estimates of the scatter in the observed relation constrain the recoil prefactor velocity to @xmath123 , which is consistent with the general relativitistic calculations of @xcite and @xcite . is there any empirical evidence for the kind of black hole processes present in our model ? @xcite have recently reported the discovery of a black hole in ngc 4435 whose mass is smaller than about 20% of the value expected from the relation . black holes with a smaller than expected mass arise naturally in our model , although masses as extreme as this could be rare . for example , in our model with @xmath47 , only @xmath124 of black holes have a mass that deviates as much or more from the mean relation as that of the black hole in ngc 4435 . however , if @xmath98 this fraction raises to @xmath125 . indirect evidence for black holes with masses below those expected from the relation has been presented by @xcite . they argue that the compact x - ray sources often seen near the centre of elliptical and spiral galaxies could be black holes of mass @xmath126 . interestingly , these sources are often displaced from the centre by a few hundred parsecs . similarly , @xcite have discovered a group of off - centre compact x - ray sources in the merger remnant ngc 3256 which , they argue , could be intermediate - mass black holes . in our model , these objects might be identified with the black holes of infalling satellites or with recently merged black holes that have been kicked out of the galactic centre , their growth stunted as a result . to investigate the spatial distribution of black holes in milky - way like galactic halos we used a set of n - body simulations to track both the satellite galaxies that host black holes and also the black holes that are ejected from their host galaxies . we find that the black hole mass function is bimodal , being composed of two overlapping populations . the lower mass population consists of black holes at the centre of orbiting low mass satellites that have not undergone recent mergers , while the slightly higher mass population is composed of wandering black holes that have been ejected from mergers that formed the central galaxies and the more massive satellites . among the latter population , we find a few supermassive ( @xmath127 ) black holes that were ejected from the main progenitor of the central galaxy sufficiently early such that the bulge of the central galaxy has had enough time to regrow a sizable black hole . there is also an intergalactic population consisting of black holes whose recoil velocity was large enough not only to unbind them from their host galaxy but also from its halo . in the future it may be possible to detect the formation of wandering black holes directly using instruments such as _ _ lisa _ _ that can directly measure the gravitational radiation emitted in the black hole black hole merger that ejects the remnant . in the meantime , detecting these black holes presents an interesting challenge . in practice , an ejected intergalactic or intrahalo black hole is likely to bring along a small cusp of stars tightly bound to it as it escapes the halo s potential . these stars would provide the only measurable way of detecting such black holes . most likely , there would not be any gas that could be accreted and hence the black holes would not be observable in the conventional way . however , as the orbits of the stars which the black hole brought with it from the bulge decay , the stars may plunge into the black hole . in addition to emitting a burst of gravitational radiation , this type of infall would tidally disrupt the star and create a small accretion disc which would radiate according to the standard physics of accretion discs . the gravitational radiation signature and accretion disc emission would provide the best way to identify these black holes . @xcite have recently discovered a quasar , he0450 - 2958 , which has no visible host galaxy and , at first sight , is a good candidate for an escaped black hole . @xcite have indeed suggested that the quasar may have been ejected from a nearby ultraluminous infrared galaxy ( ulirg ) either through the gravitational recoil process we are considering here or through a gravitational slingshot associated with three or more black holes involved in the merger responsible for the ulirg . however , @xcite have argued that the quasar is much too far from the companion galaxy to have been ejected with a velocity of @xmath51 and so favour a slingshot mechanism . ultraluminous x - ray sources ( ulxs ) or micro - quasars have long been regarded as candidates for intermediate - mass black holes , @xmath128 , radiating near the eddington limit ( see e.g. @xcite and references therein ) . this is exactly the kind of object that would be naturally identified with the orbiting intrahalo black holes predicted by our model . this interpretation , however , has two difficulties . firstly , ulxs tend to be associated with star - forming regions and their frequency seems to be correlated with the galactic star formation rate ( see e.g. @xcite ; @xcite and references therein ) . these facts lend support to the view that ulxs are stellar mass black holes emitting beamed radiation at highly super - eddington rates . the second arguement against identifying ulxs with the intrahalo black holes in our model is the difficulty of finding a suitable source of material for the black hole to accrete . there are , however , some examples of ulx s , most notably the ulx in m82 ( @xcite ) , that are probably much too bright to be explained even as exotic super - eddington luminosity stellar mass black holes ( @xcite ) . @xcite call objects like this `` hyperluminous x - ray sources '' ( hlx ) and argue that these objects are precisely the intrahalo black holes associated with satellites in our model which switch on when they come close to the galactic centre . since the exact mechanism by which these black holes would be activated is uncertain , we can not predict how common this phenomenon should be . however , our model contains , in principle , a plentiful supply of intermediate mass black holes orbiting in the halo of every galaxy which is sufficient to account for the presence of a few ulxs or hlxs in most galaxies . baugh c. m. , cole s. , frenk c. s. , lacey c. g. , 1998 , apj , 498 , 504 benson a. j. , lacey c. g. , baugh c. m. , cole s. , frenk c. s. , 2002 , mnras , 333 , 156 benson a. j. , frenk c. s. , baugh c. m. , cole s. , lacey c. g. , 2003 , mnras , 343 , 679 binney , j. & tremaine , s. , 1987 galactic dynamics . princeton university press , new jersey . blanchet l. , qusailah m. s. s. , will c. m. , 2005 apj accepted bower r. g. , benson a. j. , malbon r. , helly j. c. , frenk c. s. , baugh c. m. , cole s. , lacey c. g. , submitted to mnras , preprint ( astro - ph/0511338 ) cattaneo a. , haenhelt m. , rees m. j. , 1999 mnras 308,77 coccato , l. , sarzi , m. , pizzella , a. , corsini , e.m . , dall bonta , e. & bertola , f. , 2005 , submitted to mnras ( astro - ph/0511696 ) . cole , s. , lacey , c. g. , baugh , c. m. , frenk , c. s. 2000 mnras , 319 , 168 colbert e. , ptak a. , 2002 apjs 143 , 25 colbert e. , mushotzky r. , 1999 apj , 519 , 89 croton d. j. , springel v. , white s. d. m. , lucia g. de , frenk c. s. , gao l. , jenkins a. , kauffmann g. , navarro j. f. , yoshida n. 2005 mnras , in press , preprint ( astro - ph/0508046 ) davis m. , efstathious g. , frenk c. s. , white s. d. m. , 1985 , apj , 292 , 371 di matteo t. , springel v. , hernquist l. , 2005 natur . , 433 , 604 efstathiou , g. , lake , g. , negroponte , j. 1982 mnras , 199 , 1069 enoki m. , inoue k. t. , nagashima m. , sugiyama n. , 2004 apj , 615 , 19 fabbiano g. , 1989 ara&a , 27 , 87 favata , m. , hughes , s. a. , holz , d. e. 2004 apj , 607 , l5 ferrarese , l. & merritt , d. 2000 apj 539,l9 fitchett , m. j. 1983 mnras , 203 , 1049 gao l. , white s. d. m. , jenkins a. stoehr f. , springel v. , 2004 mnras 355 , 819 gebhardt , k. , bender , r. , bower g. , et al , s. 2000 apj , 539 , l13 gebhardt k. , rich r. m. , ho l. c. , 2002 apj 578 , l41 ghigna s. , moore b. , governato f. , lake g. , quinn t. , & stadel , j. 2000 apj 544,616 ghigna , s. , moore b. , governato f. , lake g. , quinn t. , stadel j. 1998 mnras,300,146 graham a. w. , erwin p. , caon n. , trujillo i. , 2001 apj , 563 , l11 hrring n. , rix h. w. , 2004 apj , 604 , l89 haehnelt m. g. , rees m. j. 1993 mnras 263 , 178 haehnelt m. g. , davies , m. b. , rees m. j. 2005 mnras in press ( astro - ph/0511245 ) harker , g. , cole , s. m. , helly , j. h. frenk , c. s. & jenkins , a.r . 2005 , submitted to mnras , ( astro - ph/0510488 ) hayashi e. , navarro j. f. , power c. , jenkins a. , frenk c. s. , white s. d. m. , springel v. , stadel j. , quinn t. , 2004 , mnras , 355 , 794 helly j. , cole s. , frenk c. s. , baugh c. m. , benson a. , lacey c. g. , 2003 , mnras , 338 , 903 hernquist , l. 1990 apj , 356 , 359 hoffman l , loeb a. , submitted to nature . preprint ( astro - ph/0511242 ) kaaret , p. , prestwich , a. h. , zezas , a. , murray , s. s. , kim , d .- w . , kilgard , r. e. , schlegel , e. m. , ward , m. j. 2001 , mnras 321 , 29 kaaret p. , alonso - herrero a. , gallagher j. s. , fabbiano g. , zezas a. , rieke m. j. , 2004 mnras , 348 , 28 kauffman g. , haehnelt m. , 2000 mnras , 311 , 576 king , a. r. & dehnen , w. 2005 , mnras , 357 , 275 king a. r. , davies m. b. , ward m. j. , fabbiano g. , elvis m. , 2001 , apj , 552 , l109 kormendy , j. , & richstone , d. o. 1995 , ara&a , 33 , 581 magain p. , letawe g. , courbin f. , jablonka p. , jahnke k. , meylan g.,wisotzki l. , 2005 nature , 437 , 381 lacey , c. g. , cole , s. 1993 mnras , 262 , 627 libeskind n. i. , frenk c. s. , cole s. , helly j. c. , jenkins a. , navarro j. f. , power c. , 2005 mnras , 363 , 146 madau , p. , & quataert 2004 , apjl , 606 , l17 magorrian , j. , tremaine , s. , richstone , et al . 1998 , aj , 115 , 2285 marconi , a. , & hunt , l. k. 2003 apj , 589 , l21 malbon r. , in prep mclure , r. j. , & dunlop , j. s. 2002 , mnras , 331 , 795 menou k. , haiman z. , narayanan v. k. , 2001 apj , 558 , 535 mo , s. , mao , h. & white , s.d.m . 1998 , mnras , 295 , 319 merritt , d. , & ferrarese , l. 2001 mnras 320 l30 merritt , d. , milosavlejevic , m. , favata , m. , hughes , s. a. , & holz , d. , e. 2004 apjl , 607 , l9 milosavljevic m. , merritt , d. , 2001 apj , 563 , 34 mushotzky , r. , 2004 prog . 155 , 27 - 4 navarro j. f. , frenk c. s. , white s. d. m. 1996 apj , 462,563 navarro j. f. , frenk c. s. , white s. d. m. 1997 apj , 490,493 navarro j. f. , hayashi e. , power c. , jenkins a. , frenk c. s. , white s. d. m. , springel v. , stadel j. , quinn t. , 2004 , mnras , 349 , 1039 neff s. g. , ulvestad j. s. , campion s. d. 2003 apj 599 , 1043 ostriker j. p. , 2000 , phrvl , 84 , 5258o peres a. , 1962 , phrv , 128 , 2471 peters p. c. , matthews j. , 1963 , ph rv , 13,1 435 power c. , navarro j. f. , frenk c. s. , jenkins a. , white s. d. m. , springel v. , stadel j. , quinn t. , 2003 , mnras , 338 , 14 saslaw w. c. , valtonen m. j. , aarseth , s. j. , 1974 apj 190 , 253 silk j. , rees m. j. , 1998 , a&a , 331 , l1 soltan , a. 1982 , mnras , 200 , 115 springel , v. , yoshida , n. & white , s. d. m. 2001a _ new astronomy _ , 6 , 51 springel , v. , white , s. d. m. , tormen , g. & kauffmann , g. 2001b mnras , 328 , 726 tasca l. a. m. , white s. d. m , submitted to mnras . preprint ( astro - ph/0507249 ) tremaine , s. , gebhardt , k. , bender , r. et al 2002 apj , 574 , 740 volonteri m. , haardt f. , madau p. , 2003 apj 582 , 559 ward m. 2005 , to appear in meurs e. j. a. , fabbiano g. , eds , proc . 230 populations of high energy sources in galaxies	galactic bulges are known to harbour central black holes whose mass is tightly correlated with the stellar mass and velocity dispersion of the bulge . in a hierarchical universe , galaxies are built up through successive mergers of subgalactic units , a process that is accompanied by the amalgamation of bulges and the likely coalescence of galactocentric black holes . in these mergers , the beaming of gravitational radiation during the plunge phase of the black hole collision can impart a linear momentum kick or `` gravitational recoil '' to the remnant . if large enough , this kick will eject the remnant from the galaxy entirely and populate intergalactic space with wandering black holes . using a semi - analytic model of galaxy formation , we investigate the effect of black hole ejections on the scatter of the relation between black hole and bulge mass . we find that while not the dominant source of the measured scatter , they do provide a significant contribution and may be used to set a constraint , @xmath0 , on the typical kick velocity , in agreement with values found from general relativistic calculations . even for the more modest kick velocities implied by these calculations , we find that a substantial number of central black holes are ejected from the progenitors of present day galaxies , giving rise to a population of wandering intrahalo and intergalactic black holes whose distribution we investigate in high - resolution n - body simulations of milk - way mass halos . we find that intergalactic black holes make up only @xmath1 of the total galactic black hole mass but , within a halo , wandering black holes can contribute up to about half of the total black hole mass orbiting the central galaxy . intrahalo black holes offer a natural explanation for the compact x - ray sources often seen near the centres of galaxies and for the hyperluminous non - central x - ray source in m82 .
supermassive black holes ( smbh ) are thought to dwell at the centers of most galaxies , ( kormendy @xmath7 richstone 1995 ) with masses between 10@xmath8 m @xmath9 . in principle , the abundance of smbhs today can be explained if they grow through mergers and early accretion ( schneider et al . 2002 ) , from a gaseous disk . the most likely candidates for smbh seeds are black holes that form as remnants of population iii stars at redshifts z@xmath1012 - 20 ( heger et al . 2003 , volonteri et al . 2003 , islam et al . 2003 , wise @xmath7 abel 2005 ) . these relic seeds are predicted to form in the centers of dark matter halos ( dmh ) , and have masses @xmath11 ( abel et al . 2000 , 2002 ) . dmhs form in the early universe and hierarchically merge into larger bound objects . as dmhs merge into massive halos , the seed black holes sink to the center through dynamical friction and eventually coalesce . although the seed formation stops at z@xmath112 as population iii supernovae rates drop to zero ( wise @xmath7 abel 2005 ) , smbh growth continues as dmh mergers proceed to low redshifts . from a combination of gas accretion and binary black hole coalescence , seeds can grow to intermediate mass black holes ( imbhs , with masses @xmath12 m @xmath13 ) . with continued mergers and gas accretion , it is thought that these imbhs may form the smbhs we observe today . the detection of imbhs is a matter of debate , but possible candidates are ultraluminous x - ray sources in young star - forming regions ( fabbiano 1989 , roberts @xmath7 warwick 2000 , ptak @xmath7 colbert 2004 , fabbiano @xmath7 white 2006 ) and nearby extragalactic star clusters . although this scenario works in general , real detailed understanding of how seed black holes grow is a remaining challenge . for example , there are still debates on the seed black hole mass , on the mass of halos that can host seeds , on the seed formation redshift , on the type and efficiency of gas accretion , and on the dynamics of the black hole mergers , all of which compound to yield a huge range in black hole merger rate estimates . recent black hole merger rate calculations span over three orders of magnitude : haehnelt ( 1994 ) calculates the merger rate to be r@xmath1 0.1 - 100 events per year ; menou et al . ( 2001 ) , r@xmath1 1 - 100 yr@xmath4 ; wyithe @xmath7 loeb ( 2003 ) , r@xmath1 15 - 350 yr@xmath4 ; sesana et al . ( 2004 ) , r@xmath1 10 yr@xmath4 ; and rhook @xmath7 wyithe ( 2005 ) , r@xmath1 15 yr@xmath4 . there are two general approaches to the problem : direct cosmological n - body simulations and analytical techniques based on a press - schechter ( ps ) formalism ( press @xmath7 schechter 1974 ) . both methods extract merger rates from the dmhs merger tree . n - body simulations have the advantage that the evolution of density fluctuations is followed in complete generality , without the need for any of the assumptions involved in creating ps merger trees . extended press - schechter theory ( eps ) combines the ps halo mass function with halo merger rates derived by lacey @xmath7 cole ( 1993 ) , and it stands as the most widely used method for calculating merger rates . unfortunately , eps is mathematically inconsistent ( erickcek et al . 2006 ) since it provides two equally valid merger rates for the same pair of dmhs . it is also unclear whether eps reproduces the dmh mass function at high redshifts ( reed et al . 2007 ) . with n - body simulations , there are , in principle , no constraints on the halo mass , structure or kinematics . unfortunately , the mass and spatial resolution needs to be extremely high , and even in the best resolved n - body simulations of a large volume @xmath14cdm universe ( nagashima et al . 2005 ) , the mass resolution sets the minimum dmh mass to 3@xmath1510@xmath16@xmath17 for halos with as few as ten particles . hence using direct n - body simulations to track the dynamics of seed black holes as they grow within a large comoving volume is out of reach of our current technology . in this paper , we have developed a hybrid method to follow the merger history of seed black holes as they grow . we performed a high resolution cosmological n - body simulation in the unexplored parameter space of a _ small cosmological volume _ but with very high mass resolution to achieve well - resolved halos with a minimum of 32 particles and mass as low as m@xmath18=2.8@xmath19 . the goal was to look at a representative `` local group '' structure , comparable to the one which hosts the milky way , and to resolve as low a mass as feasible . we improved the algorithms for identifying dark matter structures and developed set of physically - motivated criteria for seeding dmhs . we then constructed the dmh merger tree , which provides a testbed to study the effects of different gas accretion scenarios . in addition to the merger rate , we extracted observables such as the black hole mass function over cosmological time . this method is a major improvement in calculations of mbh merger rates and presents the first step toward a full treatment of the black hole growth problem . massive black holes mergers will be one of the prime signals for future space - based gravitational wave observatories like lisa ( laser interferometer space antenna ) and bbo ( big bang observer ) . possible detection depends on a number of parameters : the mass ratio of merging black holes , the total mass of the black hole binary and the redshift . we describe our simulation and dmh seeding criteria in section 2 ; obtaining dmh and black hole merger trees in section 3 and 4 together with description of black hole growth models ; black hole merger rates in section 5 ; finally , we discuss our results in section 6 . in our numerical simulations , we use gadget ( springel et al . 2001 ) to evolve a comoving 10 mpc@xmath20 section of a @xmath14cdm universe ( @xmath21=0.3 , @xmath22=0.7 and h=0.7 ) from @xmath23 to @xmath24 . we refine a sphere of 2 mpc in the box to simulate at a higher resolution with 4.9@xmath25 high - resolution particles ( softening length 2 kpc comoving ) . the rest of the box has 2.0@xmath25 low - resolution particles ( softening length 4 kpc comoving ) . the mass of each high resolution particle in this simulation is 8.85@xmath26m@xmath27 , and the mass of each low - resolution particle is 5.66@xmath28m@xmath27 . a detailed description of this simulation can be found in micic et al . 2006 . identifying and following a real dmh in a cosmological simulation is technically challenging . if the minimum number of particles in a halo is set too low , the dmh identified at one time - step might later disperse . in the widely used friends - of - friends algorithm ( fof ) , if the linking length ( b ) is too large , large dark matter structures might bridge through a small number of tidally stripped particles . when this occurs , the fof algorithm can not properly identify dmhs . a linking length of 0.2 has traditionally been used ( davis et al . 1985 ) since the dmh mass function obtained with it proved to be consistent with observations at low redshifts and mass functions obtained by ps theory . however , very little is known about dmh mass functions at high redshifts z@xmath1010 and low masses m@xmath610@xmath29@xmath17 , or about the validity of this linking length . for example , gao et al . ( 2005 ) noted that the first massive halos are found in regions containing many smaller halos which line up along filaments and sheets and can not be identified by the fof algorithm when a linking length of b=0.2 is used . the only criterion for initializing linking length comes from the spatial resolution in the performed numerical simulation . the linking length has to be larger than the softening length . high resolution simulations have small softening lengths , which allow the linking length to be smaller as well . consequently , in our high resolution simulation , we set b=0.1 . with this value for linking length we eliminated all the problems in identification of dark matter structures . we used p - groupfinder algorithm ( springel 2000 ) to identify dmhs in each of our 104 gadget output files . the analysis starts with a fof algorithm ( b=0.1 ) which spatially separates dark matter halos . the minimum number of particles per halo is set to 32 . the minimum dmh mass resolved with 32 particles in our simulation is m@xmath18=2.8@xmath19 . then , we find a gravitationally bound counterpart for each fof halo and continue with the subfind algorithm to identify substructures in every dmh . we require that halos and subhalos be virialized to be considered real . this requirement is met when the dmh density is above 200 times the average density of the universe at a specific redshift . since we are studying merger rates for pop iii massive black holes , we then select those halos that are capable of hosting pop iii stars , using primordial supernova rates ( wise @xmath7 abel 2005 ) . primordial stars can form in dmhs with a minimum mass of 4.16@xmath1510@xmath30 @xmath17 if h@xmath31 cooling is the primary mechanism and if the uv background has a negative feedback . they can also form through hydrogen atomic line cooling in halos with mass m@xmath32@xmath1010@xmath33@xmath17[(1+z)/10]@xmath34 . the lifetime of pop iii stars is @xmath13 myrs and the end product is a population of seed black holes with 100@xmath17@xmath6m@xmath35@xmath61000@xmath17 . the first pop iii seed black holes form at redshift z@xmath120 and the last pop iii stars die at redshift z@xmath112 . we select the interval 12@xmath6z@xmath619 to seed those dmhs which have masses m@xmath32@xmath1010@xmath33@xmath17[(1+z)/10]@xmath34 . at z=19 all dmhs with mass above m@xmath18=3.5@xmath19 and at z=12 all of the dmhs with mass above m@xmath18=6.7@xmath19 are seeded . starting with the first snapshot `` i '' , we identify dmhs with the set of criteria described above . the most bound particle at the center of each dmh that satisfies the criteria is set to be a seed black hole . at the following snapshot `` i+1 '' , we identify halos from the previous `` i '' snapshot . at `` i+1 '' we again seed new halos that satisfy our criteria . by `` new '' halos we consider only those halos that do not have any particles that were part of any already seeded halo . this is because once a halo is seeded it can not form a new black hole due to feedback from the uv background of pop iii stars ( machacek et al . meanwhile , we also trace the parent dmh at `` i '' for every particle in every dmh at `` i+1 '' , thus constructing a complete merger history . the method is repeated through the redshift range 12@xmath6z@xmath619 . seeding stops at redshift z=12 but we continue with creating a merger history until redshift z=0 . in total we have 1447 seed black holes . after constructing the dmh merger tree , we turn our attention to the evolution of the seeds . the n - body simulation tracks all the seed black holes and their host halos from z=19 - 0 . in order to study black hole growth as a function of time , environment , and merger type , we constructed three black hole merger models which we detail below . we parameterize the degree of gas accretion involved in growing the black hole and the efficiency of binary coalescence . in the fiducial case , we modeled black hole mergers in the simplest possible way . we assume that the hardening timescale ( the time for two black holes to form hard binary ) is rapid , that the loss cone is full ( berczik et al . 2006 , sigurdsson 2003 , holley - bockelmann @xmath7 sigurdsson 2006 ) , that no black holes are ejected , and that there is no gas accretion involved . this gives us a strawman model with which to compare more realistic black hole growth scenarios . we assign seed black hole masses of 200@xmath17 . when two dmhs merge , their seeds form a binary and merge at the center of the remnant with the total combined mass of its progenitors . if more than two dmhs merge , as often happens early on , the merger is sorted by dmh mass , since dynamical friction is stronger on higher masses . escala et al . 2006 suggests that the black hole binary will merge within a few times 10@xmath36 yr . since dmhs typically merge in 10@xmath33 yr , the mbhs will coalesce soon after the dmhs merge in this model . as the halos evolve to the present day , black holes can only grow in mass through mergers , and in our ( small cosmological volume ) simulation reached a maximum mass of m@xmath35=3@xmath37 . we find that this value is too small to match observations . for a sustained eddington accretion , the mass growth rate is : @xmath38 where t@xmath39 is the salpeter time - scale , t@xmath39@xmath14@xmath28 yr from the recent observations ( hu et al . 2006 ) . in our models , we assume that accretion is triggered by galaxy mergers and lasts @xmath40t@xmath39 . during this time , the black holes approximately double their mass . however , we do not assume that every merger supplies enough gas for the black hole to double its mass . for example , during minor galaxy mergers ( mass ratios much larger than 10:1 ) , the more massive galaxy will shred the satellite ( e.g. white 1983 , holley - bockelmann @xmath7 richstone 1999 ) , and the black hole will sink toward the central black hole without a large reservoir of gas . on the other hand , the merger of comparable ( mass ratios less than 10:1 ) galaxies will generate ample gas accretion onto the central black hole , as the satellite triggers a central starburst ( mihos @xmath7 hernquist 1994 ) . here we distinguish two cases , depending on the mass ratio of merging dmhs . the first is a more conservative criterion that allows black holes to accrete gas if the mass ratio of host dmhs is less than 4:1 ( 4:1 accretion ) . the second case sets an upper constraint on the mass of the final black hole by allowing seeds to accrete gas as long as the merging dmhs have a mass ratio less than 10:1 ( 10:1 accretion ) . in the following section , we compare all three growth scenarios : black hole growth through mergers only ( dry growth ) ; growth through mergers combined with 4:1 gas accretion ( 4:1 growth ) ; and finally , growth through mergers combined with 10:1 gas accretion ( 10:1 growth ) . these are simplifying assumptions made to explore parametrically the range of solutions in our scenario . from z=19 to z=12 , we identified all of the dmhs in the manner described in the previous section . with this approach , we obtained the initial positions and formation redshifts for 1477 seed black holes , approximately 100 new mbhs per snapshot . we assigned an initial mass of 200@xmath17 to each mbhs and traced their merger history from 0@xmath6z@xmath619 . throughout the simulation , one dmh emerges as the largest in mass and dominates the dynamics of its group . its mass at z=0 is m@xmath41=4.5@xmath1510@xmath42@xmath17 which roughly corresponds to andromeda s dmh , and is somewhat higher than estimates for the milky way s halo ( dehnen et al 2006 ) . this halo grows through the smooth accretion of dark matter particles as well as dmh mergers . meanwhile , the central black hole grows through gas accretion ( if we included it ) and mergers . we traced the growth of both the primary halo and its massive black hole as a function of redshift ( figure 1 ) . figure 1 compares the growth of both the black hole ( dry ) and its host dmh . overplotted is a descriptive redshift dependence because it exists only where the curve has a negative slope . initially , the primary halo grows in mass only through smooth accretion and reaches m@xmath41=10@xmath43 by z=15 ( fig . 1 ) . during this period , the black hole remains at its initial mass . at all other redshifts , the growth of primary halo and any other dmh can be described as a cycle of steady accretion of the surrounding dark matter punctuated by rapid growth through mergers with incoming dmhs . we present the dmh merger rates in figure 2 . in comparing figure 1 and figure 2 , notice that changes in merger rate catalyze both the mbh and primary halo mass growth . figure 2 presents merger rates as the number of mergers per unit time per unit redshift for all halos , while figure 1 traces the merger history for the primary halo only . nevertheless , since the primary halo participates in 90@xmath44 of all dmh mergers , figure 2 essentially follows the evolution of the primary halo . at z@xmath111 , the central black hole with mass m@xmath45 has reached the imbh range . at this redshift , the merger rate reaches maximum ( fig 2 . ) . consequently , the imbh grows quickly from @xmath46 to @xmath47 although surprisingly the bound dmh mass decreases.z@xmath612 ( cf figure 2 ) , increases the number of unbound dark matter particles , since there is not enough time for reaccrete before another merger . ] merger rates decrease at z@xmath610 but experience two more peaks with periods of violent mergers around z=6 and z=3 ( figure 2 ) . at z=0 , there is one smbh with mass m@xmath0=2.9@xmath1510@xmath48@xmath17 at the center of the primary halo ( figure 1 and figure 3 thick line ) . this mass is too small to match present day observations , so we know , as has been stated , that mergers alone can not create a smbh . as described before , we used the salpeter approximation to model different gas accretion scenarios where the accretion efficiency depends on the mass ratio of merging dmhs . in all three scenarios , the black hole reaches the imbh range at z@xmath112 ( figure 3 ) . as the number of dmh mergers increases , the imbh continues growing faster , naturally , if gas accretion is more efficient . in a 4:1 growth scenario , the imbh grows to a @xmath49 smbh by z=5.5 . at these redshifts , low mass ratio mergers of dmhs are depleted and smbh growth through gas accretion stops . therefore , the duty cycle of the active galactic nucleus ( agn ) hosting this smbh dies at z=5.5 , too . in this model , the smbh mass remains essentially constant for @xmath65.5 , because it only experiences very high mass ratio mergers with 200 - 1000@xmath17 black holes . finally , at z=0 the agn has evolved into galaxy hosting a @xmath49 smbh at its center , strikingly similar to the mass of the milky way smbh ( schodel et al . 2002 , ghez et al . 2003 , 2005 ) . a highly efficient 10:1 growth scenario yields an upper constraint of @xmath50 comparable to the smbh observed in m31 ( bender et al . 2005 ) . figure 4 shows the smbh mass as a function of central velocity dispersion and redshift for our three growth scenarios . we obtained the velocity dispersion for dark matter at 10kpc comoving from the halo center . recall that 10kpc comoving corresponds to 700pc at z=19 and 14 kpc at z=0 . similar to figure 3 , every decrease in dark matter velocity dispersion is a consequence of high merger rates at a specific redshift . the maximum value for the velocity dispersion obtains at z=0 , @xmath51=120@xmath52 . when compared to the m-@xmath51 relation from gebhardt et al . 2001 and merritt @xmath7 ferrarese 2001 , a smbh of @xmath13@xmath19 has a stellar velocity dispersion of @xmath51=120@xmath52 at z=0 , which matches the smbh mass in the 10:1 growth scenario . of course our m-@xmath51 relation is derived only from the dark matter and at much larger radii , but this may give us a hint that the m-@xmath51 relation is tied to the dynamics of the global potential . exploring the formation of more massive smbhs under this set of assumptions , including the redshift at which growth occurs and the merger rate as a function of redshift is beyond the scope of this paper . although there are no studies of agn duty cycle for z@xmath536 and m@xmath0@xmath54 ( wang et al . 2006 ) , it is probable that the agn fueling mechanism gas accretion is similar to that of higher mass galaxies . the smbh in 4:1 and 10:1 growth scenarios grows to its final mass of m@xmath0 @xmath55 - @xmath56 in the redshift range 5.5@xmath6z@xmath66 . as major mergers dictate the growth at z@xmath536 , the smbh evolves as a low luminosity agn with the duty cycle governed by on and off switching of accretion onto the black hole . the agn duty cycle drops to zero at z@xmath16 as the host galaxy becomes too large for major mergers to occur and both the smbh and its host galaxy enter a quiet phase of their growth through mergers with much smaller galaxies . interestingly , all incoming small galaxies at z@xmath16 in our model carry an imbh @xmath11000@xmath17 . this sets the stage for a phase where each new galaxy merger can be characterized as a dwarf starburst galaxy with a central imbh accreting gas while sinking toward the smbh at the center . in other words , in these later phases the lower mass incoming black hole may accrete gas , but we assume no significant gas accretion onto the central smbh because of the high mass ratio of the black holes . with the exception of the largest black hole in our simulation , all other black holes fall into the imbh range : 200@xmath57m@xmath5810@xmath59 . these black holes are of the special interest as candidates for ulx sources ( mii @xmath7 totani 2005 ) . ulxs are interpreted as massive black holes accreting gas in starburst galaxies , although this is still a matter of debate ( king et al . 2001 ) . figure 5 shows the merger rates for different mass ratio ranges and total binary masses . when deconstructed into ranges of mass ratios and log total binary mass ( defined as p = log(m@xmath60 + m@xmath31 ) ) , these merger rates can be used to predict the number of ulx sources in starburst galaxies throughout the galaxy evolution , assuming that merging galaxies are reasonably gas rich . for the mass ratio 1@xmath6m@xmath60/m@xmath31@xmath6110 , we see only minor differences between the three growth scenarios for any range of binary mass ( fig 5a , 5b , 5c ) . for z@xmath1010 black holes grow mostly through mergers . for z@xmath18 , mass ratio @xmath6110 and binary mass @xmath62m@xmath63 ( 3@xmath6p@xmath614 ) the merger rate is r@xmath110 yr@xmath4 in all three growth scenario and it reaches a maximum r@xmath3@xmath130 yr@xmath4 for p@xmath613 . at lower redshifts the merger rates decrease rapidly ( fig 5a , 5b , 5c ) , and gas accretion becomes important ( figure 3 ) . nevertheless , there are no changes in fig 5a , 5b , 5c for z@xmath6110 since the mass ratio of merging black holes becomes very large as the higher mass black holes gain mass preferentially ( figure 6 ) . in our scenario , only the black holes at the centers of merging dmhs grow through gas accretion ; a large number of dmhs remain isolated for most of the simulation and merge late in high mass ratio encounters ( middle and bottom panels of fig 5 ) . the difference between the top and bottom panels in figure 5 shows that most of the black hole mergers at high redshifts are with 1@xmath6m@xmath60/m@xmath31@xmath6110 mass ratio . since black holes grow faster with more efficient gas accretion , the merger rates for mass ratio 10@xmath6m@xmath60/m@xmath31@xmath61100 and total binary mass @xmath64m@xmath65 , shift toward higher redshifts in our accretion models ( fig 5d , 5e , 5f ) . in this mass ratio range , 3@xmath6p@xmath614 mergers peak at 8@xmath10z@xmath66 , with a merger rate of @xmath110 depending on the growth scenario . very high mass ratio mergers @xmath67m@xmath60/m@xmath68 are of the special interest ( fig 5 . bottom panels ) . at z@xmath616 , high mass ratio dmh mergers correspond to a dwarf galaxy or globular cluster being consumed by massive galaxy . the imbh carried by smaller counterpart will eventually coalesce with smbh at the center by first forming a binary with total mass p@xmath535 and high mass ratio @xmath67m@xmath60/m@xmath68 . this is an ideal lisa source , as will be explored in more detail in our next paper . as a result , the merger rates in figure 5d , 5e , 5f and figure 7 for the above total binary mass and mass ratio ranges directly correspond to the number of imbhs per galaxy per redshift . since the accretion timescale is approximately half the merger time scale , the ulx number density is n@xmath5=r/2 . in the case of dry growth ( fig 5 g ) , a black hole binary with mass @xmath69m@xmath65 has an approximately constant and high merger rate of r@xmath115 yr@xmath70 for 2@xmath61z@xmath616 . this corresponds to @xmath17 ulx sources per starburst galaxy at 2@xmath61z@xmath616 . similarly the @xmath69m@xmath65 binary in the 4:1 growth scenario has r@xmath610 yr@xmath4 ( fig 5h ) and n@xmath5=5 . more importantly figure 5h shows n@xmath5@xmath532 at 2@xmath61z@xmath6110 . for even higher mass ratios @xmath71m@xmath60/m@xmath72 , the merger rate is r@xmath612 yr@xmath70 for z@xmath63 . figure 7 shows merger rates for the 10:1 growth scenario and high mass ratio @xmath71m@xmath60/m@xmath73 for @xmath74m@xmath75 mergers . ulx sources in figure 7 and figure 5 g , 5h have constant number density of @xmath110 per starburst galaxy at z@xmath76 . the ulx number density drops rapidly at z=2 . note that these ulx sources are not the same as the ulxb sources seen in local universe with much shorter accretion lifetimes . we compare our results with those obtained by various black hole merger models from eps theory . figure 8 shows the merger rates for four press - schechter models decribed in sesana et al . 2007 . in the vhm model , massive dmhs ( m@xmath77=10@xmath78 - 10@xmath79 @xmath17 ) are seeded with m@xmath35@xmath1200@xmath17 black holes at z=20 ; in the kbd model , low mass halos ( m@xmath77=10@xmath30 - 10@xmath36 @xmath17 ) are seeded with m@xmath35@xmath15@xmath80 at 15@xmath6z@xmath620 ; and the bvr models explore different redshift ranges for seeding black holes in halos : m@xmath35=@xmath81 at 15@xmath6z@xmath620 in the bvrhf model and 18@xmath6z@xmath620 in the bvrlf model . figure 8a shows superimposes their merger rates with our results . since our numerical simulations have smaller mass resolution than the vhm model , their dmhs ( and therefore black holes ) start merging later with a merger rate peaking at z=12 . our maximum merger rate is also higher because our range of seeding redshifts is larger 12@xmath6z@xmath619 . the mergers in question are small mass ratio mergers , with m@xmath82 ( figure 8b ) . at z@xmath610 our merger rates match the kbd model . however , we have multiple peaks at z=6 and z=3 from episodes of violent dmh dynamics that is not depicted by press - schechter theory ( figure 8a ) . this is an advantage of numerical simulations , one that acts to increase the merger rate , the black hole growth rate , and the mass ratio of typical mergers drastically . in order to achieve @xmath83m@xmath84 ( figure 8c ) with press - schechter models , one must start at higher redshifts and assume higher initial black hole masses . the kbd model starts with high mass seeds , and as a result the kbd merger rates peak for @xmath83m@xmath84 , and are larger than the n - body rates , both in this mass range and at z@xmath535 . we `` spend '' bh seeds naturally during mergers so in this mass range our peak is m@xmath82 at z@xmath535 . we used a high - resolution cosmological n - body simulation to study the formation and growth of seed black holes into smbhs and derived black hole merger rates . we used physically - motivated formalisms for seeding dmhs with black holes . better understanding of the initial mass function for population iii black holes will improve the accuracy of our results in the future . of course , black hole seeds do not necessarily have to be population iii black holes . we can use our approach to test different seeding scenarios . for example , it can be used to set constraints for primordial black hole formation ( mack et al . 2006 ) . depending on the assumed growth scenario , we showed that gas accretion combined with hierarchical mergers of massive black holes leads to the formation of a sagittarius a * type black hole in the center of a milky way - sized dmh . in the case of very efficient gas accretion , an m31-sized smbh can form . in both cases , the smbh reaches its final mass at z@xmath16 . we showed that z=6 may be a critical redshift for the transition from the agn duty cycle dominated by high mass ratio dmh mergers to a starburst galaxy phase where low mass ratio dmh mergers supply a galaxy with a constant population of ulxs up to z=2 . we argue that the z=6 turning point is consistent with the maximum of agn luminosity function at z=2 since the smbhs in our simulation have masses and luminosities below the currently observable range . also note that potential ulx sources in our simulation are at z@xmath532 . they may correspond to nuclear clusters at the centers of dwarf galaxies , and therefore likely sites of imbhs formation . this population is different from the low redshift ulx population which may form from globular cluster dynamics and are likely undergoing short lived binary accretion . the final stages of the black hole merger are followed by an emission of gravitational radiation , low in frequency but relatively high in amplitude . in one of our follow up papers , for each binary black hole inspiral and merger , the expected gravitational wave signal for the lisa will be determined , and the lisa event rate as a function of time calculated . in particular , we will study whether lisa observations will be able to distinguish between different assembly scenarios . one interesting source we will be able to constrain will be imbh / smbh mergers . we will calculate lisa detectability of imbh mergers for different mass ranges and binary mass ratios . at this point , rough estimates can be made from figure 4 , considering the known ranges for lisa sensitivity . in case of dry growth ( fig 4 g ) , a black hole binary with mass @xmath69m@xmath65 is in the lisa range with a very high merger rate of @xmath115 yr@xmath70 . interestingly , just as smbhs dim electromagnetically at z=6 , they turn on in the gravitational wavebands . there are a number of processes that might suppress massive black hole merger rates . we assumed that every first star will produce black hole as opposed to neutron star or pair detonation with no remnant . we also assume that black holes merge efficiently and that recoil ejection which is a function of spin , orientation , eccentricity , and mass ratio of merging black holes is negligible . all of these processes will be addressed in the follow up paper . finally , since our @xmath85 black hole was in place so early , it might be tempting to interpret our results as an indictment against downsizing , the recent observation that the number density of low mass agns peaks at @xmath86 ( cowie et al . 2003 , merloni 2004 , heckman et al . . however , we caution that in order to achieve such high mass resolution , we simulated a small volume , and are therefore plagued by small number statistics . in fact , if we were to scale our simulation up in mass , so that the most massive smbhs were @xmath87 , we find some indication that the more isolated dmhs will host low mass smbhs that gain most of their mass at late redshifts compared to the most massive smbh . we plan to explore the growth of black holes as a function of environment and smooth accretion efficiency in a follow up paper . we would like to thank nasa s columbia high end computing program for a generous time allocation , and the center for gravitational wave physics at pennsylvania state university for sponsoring this research . this research was supported by a grant from the nsf , phy 02 - 03046 and from nasa s atp nng04gu99 g . ss would also like to thank kitp at ucsb and slac at stanford university for their hospitality . abel , t. , bryan , g. , & norman , m. , 2000 , apj , 540 , 39 abel , t. , bryan , g. , & norman , m. , 2002 , sci , 295 , 93a bender , r. , kormendy , j. , bower , g. , 2005 , apj , 631 , 280 berczik , p. , merritt , d. , spurzem , r. , bischof , h.p . , & 2006 , apj , 642l , 21b cowie , l.l . , barger , a.j . , bautz , m.w . , & brandt , w.n . , garmire , g.p . , 2003 , apj , 584 , l5 davis , m. , efstathiou , g. , frenk , c. , white , s.d.m . , & 1985 , apj , 292 , 371 dehnen , w. , mclaughlin , d.e . & sachania , j. , 2006 , mnras , 369 , 1688 erickcek , a.l . , kamionkowski , m. , & benson , a.j . , 2006 , mnras , tmp , 940e escala , a , larson , r.b . , coppi , p.s . , & mardones , d. , 2005 , apj , 630 , 152e fabbiano , g. , 1989 , ara&a , 27 , 87 fabbiano , g. , white , n.e . , 2006 , & in compact stellar x - ray sources , ed . w. h. g. lewin & m. van der klis & ( cambridge : cambridge univ . press ) , 475 gao , l. , white , s.d.m . , jenkins , a. , frenk , c.s . , springel , v. , & 2005 , mnras , 363 , 379 gehz , a.m. , duchene , g. , matthews , k. , 2003 , apj , 586 , 127 gehz , a.m. , salim , s. , hornstein , s.d . , 2005 , apj , 620 , 744 haehnelt , m.g . , 1994 , mnras , 269 , 199 milosavljevic , m. , apjl , submitted , astro ph/0306074 heckman , t.m . , kauffmann , g. , brinchmann , j. , & charlot , s. , tremonti , c. , white , s.d.m . , 2004 , apj , 613 , 109 heger et al . , 2003 , 2003 , apj , 591 , 288h holley - bockelmann , k. , & sigurdsson , s. , 2006 , astro - ph , 1520h holley - bockelmann , k. , & richstone , d.o . , 1999 , apj , 517 , 92h islam , r. r. , taylor , j. e. , silk , j. , & 2003 , mnras , 340 , 647i king , a.r . , davies , m.b . , ward , m.j . , fabbiano , g. & elvis , m. , 2001 , apjl , 552 , 109 kormendy , j. , richstone , d. , 1995 , ara&a , 33 , 581k lacey , c. , cole , s. , 1993 , mnras , 262 , 627 machacek , m.e . , bryan , g.l . , & abel , t. 2001 , apj , 548 , 509 mack , k.j . , ostriker , j.p . , ricotti , m. , astro - ph/0608642 menou , k. , haiman , z. , narayanan , v.k . , & 2001 , apj , 558 , 535 merloni , a. , 2004 , mnras , 353 , 1035 merritt , d. , ferrarese , l. , 2001 , apj , 547 , 140 m mihos , j.c . , hernquist , l. , 1994 , apj , 425l , 13 m mii , h. , totani , t. , 2005 , apj , 628 , 873 micic , m. , abel , t. , sigurdsson , s. , 2006 , mnras , 372 , 1540 m nagashima , m. , et al . 2005 , apj , 634 , 26n press , w.h . , schechter p. , 1974 , apj , 187 , 425 ptak , a. , colbert , e. , 2004 , apj , 606 , 291 reed , d.s . , bower , r. , frenk , c.s . , jenkins , a , theuns , t. , & 2007 , mnras , 374 , 2r rhook , k.j . , wyithe , j.s.b . , 2005 , mnras , 361 , 1145 roberts , t. , warwick , r. , 2000 , mnras , 315 , 98 sesana , a. , haardt , f. , madau p. , & volonteri , m. , 2004 , apj , 611 , 623 sesana , a. , volonteri , m. , haardt , f. , & 2007 , astro - ph/0701556 schneider , r. , ferrara , a. , & natarajan , p. , omukai , k. , 2002 , apj , 571 , 30 schodel , r. , ott , t. , genzel , r. , 2002 , nature , 419 , 694 sigurdsson , s. , 2003 , cqgra , 20s , 45s springel , v. , yoshida , n. , & white , s. d. m. , 2001 , newa , 6 , 79 springel , v. , 2000 , mpa volonteri , m. , haardt , f. , & madau , p. , 2003 , apj , 582 , 559 wang , j.m . , chen , y.m . , zhang , f. , 2006 , apj , 647l , 17w white , s.d.m . , 1983 , apj , 274 , 53w wise , j. , h. , abel , t. , 2005 , apj , 629 , 615w wyithe , j.s.b . , loeb , a. , 2003 , apj , 590 , 691 ) , and as a function of redshift . redshift dependence is descriptive hence filled circles correspond to data points in terms of redshift . @xmath51=120 @xmath52 at z=0 and it fits the m-@xmath51 relation for 10:1 growth accretion . ] + m@xmath31 ) . 1@xmath6m@xmath60/m@xmath88 in a , b , c ; 10@xmath6m@xmath60/m@xmath89 in d , e , f ; 100@xmath6m@xmath60/m@xmath90 in g , h , i ; dry growth in a , d , g ; 4:1 growth in b , e , h ; 10:1 growth in c , f , i. most of the mergers are low mass ratio mergers at z@xmath5310 . however , high mass ratio mergers at z@xmath6110 ( g , h ) are in the lisa range and have large merger rates r @xmath1 15 yr@xmath4 for a wide range of redshifts 2 @xmath61 z @xmath61 10.,width=672 ] = m@xmath60 + m@xmath31 for four models described in sesana et al . 2007 , vhm in thin black line ; kbd in dot - red line ; bvrlf in blue - dashed line and bvrhf in black - dot line . overplotted are merger rates from our n - body simulations for dry growth in thick black line ; 4:1 growth in red dashed line and 10:1 growth in blue dashed - dot line . we predict massive black hole merger rates to be larger than those predicted by semi - analitical models in the range 2 @xmath61 z @xmath61 15 . numerical simulations also depict episodes of violent dmh dynamics which can be seen in multiple peaks at z=11 ; z=6 ; and z=3.,width=672 ]	understanding how seed black holes grow into intermediate and supermassive black holes ( imbhs and smbhs , respectively ) has important implications for the duty - cycle of active galactic nuclei ( agn ) , galaxy evolution , and gravitational wave astronomy . most studies of the cosmological growth and merger history of black holes have used semianalytic models and have concentrated on smbh growth in luminous galaxies . using high resolution cosmological n - body simulations , we track the assembly of black holes over a large range of final masses from seed black holes to smbhs over widely varying dynamical histories . we used the dynamics of dark matter halos to track the evolution of seed black holes in three different gas accretion scenarios . we have found that growth of a sagittarius a * - size smbh reaches its maximum mass m@xmath0=@xmath110@xmath2 at z@xmath16 through early gaseous accretion episodes , after which it stays at near constant mass . at the same redshift , the duty - cycle of the host agn ends , hence redshift z=6 marks the transition from an agn to a starburst galaxy which eventually becomes the milky way . by tracking black hole growth as a function of time and mass , we estimate that the imbh merger rate reaches a maximum of r@xmath3=55 yr@xmath4 at z=11 . from imbh merger rates we calculate n@xmath5=7 per milky way type galaxy per redshift in redshift range 2@xmath6z@xmath66 . stars : population iii , intermediate mass black holes , supermassive black holes , gravitational waves , dark matter halos
miras are pulsating stars with initial masses between 0.8 and 8 solar masses in the asymptotic giant branch ( agb ) phase , and eject material via stellar winds into the interstellar medium ( e.g. , habing 1996 ) . the ejected material contain chemical elements that have been dredged up from the interior ( e.g. , carbon and s - process elements ) . a large amount of dust forms in the ejecta of the miras , and then such dust grains regulate the cooling of the interstellar medium and the fragmentation of collapsing molecular clouds into stars . thus , miras play an important role in providing the heavy elements and dust grains from the early universe to the present day . the lack of appropriate theoretical models of miras currently prevents us from determining their fundamental properties , such as metallicities , based on observational data for miras . therefore , miras in stellar systems with known metallicity and/or age distribution can be important tracers to study the evolution of miras and their impacts on chemical enrichment . for example , most of the galactic globular clusters are old stellar systems with a single metallicity or a narrow metallicity distribution , offering an important sample of low - mass and metal - poor miras ( frogel & whitelock 1998 ; feast et al . we note that miras are found only in the clusters with @xmath6>-1}$ ] . another interesting sample is low- to intermediate - mass miras found in the magellanic clouds . a formidable amount of literature exists on those objects ( e.g. ita et al . 2004ab ; fraser et al . 2008 ; soszyski et al . 2009 ; groenewegen et al . galactic dsphs provide us with a sample of even lower metallicity objects than the globular clusters . it is known that the fainter galaxy tends to have the lower mean metallicity ( norris et al . therefore , the faint dsphs are excellent places to study metal - poor miras if any mira is found . recent monitoring surveys in galactic dsphs have discovered several miras ( fornax dsph , whitelock et al . 2009 ; leo i dsph , menzies et al . 2010 ; sagittarius dsph , lagadec et al . 2009 ; sculptor dsph , menzies et al . 2011 ) . among the dsphs with previously known miras , the sculptor dsph is the most metal deficient one , which has the metallicity distribution with a peak at @xmath6}=-1.56 $ ] and a dispersion of 0.48 ( kirby et al . sloan et al . ( 2009 ) reported the evidence of circumstellar dust around one of the miras in the sculptor dsph ( menzies et al . 2011 ) , based on the _ spitzer space telescope _ spectroscopy . this suggests that agb stars in low - metallicity environments can make a significant contribution to dust formation in the early universe . our target galaxy , the sextans dsph , shows a metallicity distribution with a peak at @xmath6}=-1.9 $ ] and is one of the most metal - poor dsphs in the galaxy ( battaglia et al . so far , two monitoring surveys have been conducted for the center of this galaxy , revealing dozens of short - period variable stars ( mateo et al . 1995 ; lee et al . no mira has been found previously.in this paper , we report the discovery of two miras toward the sextans dsph . in section 2 , we describe optical and infrared photometric observations , and discuss their membership to the sextans dsph . in section 3 , we discuss the chemical properties and their impacts . in order to explore miras in the sextans dsph , we selected two target stars from the photometric catalogs presently available . these are listed in table [ tab : object ] . the first target # 1 , sdss [email protected] , was selected using the color criteria , @xmath8 , @xmath9 on the 2mass catalog ( skrutskie et al . 2006 ) and @xmath10 , @xmath11 on the sdss catalog ( adelman - mccarthy et al . these criteria are also used for our monitoring survey of miras in the galactic halo ( sakamoto et al . , in preparation ) . the target # 1 is carbon - rich , showing a spectrum with the strong cn absorption band at 7900@xmath12 ( mauron et al . 2004 ; cruz et al . the second target # 2 , sdss [email protected] , was later added because of its variation detected in quest1 ( quasar equatorial survey team , phase 1 ) variability survey ( rengstorf et al . the @xmath13-band light curve of # 2 over 2 years in the quest1 showed a variation with a large amplitude ( @xmath14 mag ) and a long period ( over 100 days ) , although their time sampling was not good enough to estimate the period . the target # 2 is oxygen - rich , showing a spectrum with clear tio molecular absorption lines ( suntzeff et al . 1993 ) . we conducted photometric monitoring observations of the two selected targets in the direction of the sextans using the 2kccd camera attached to the 105-cm f/3.0 schmidt telescope at kiso observatory ( itoh et al . 2001 ) . the observations started in december 2008 and february 2010 for the targets # 1 and # 2 , respectively , and were repeated until february 2012 . time series @xmath15-band images were obtained.the data were reduced following standard procedures with iraf , including bias subtraction ( both the level of the overscan region in each image and the bias pattern taken on each night ) and the flat - field correction with @xmath15-band dome - flat images . instrumental magnitudes of the targets and comparison stars were measured with aperture photometry using the iraf / apphot package . the comparison stars were selected from the sdss database ( adelman - mccarthy et al . 2008 ) , and their @xmath15 magnitudes were calculated by using the transformation of jordi et al . ( 2006 ) , @xmath16 using these @xmath15 magnitudes for calibrating the magnitude scale , we obtained the magnitudes of our target stars as listed in tables 2 and 3 . [ fig : lc ] plots the @xmath15 variations against the modified julian date ( mjd ) . both stars show the long - period and large - amplitude variation characteristic of either miras or semi - regular variables . the peak - to - valley amplitudes ( @xmath17 ) are 3.72 mag for the target # 1 and 0.94 mag for the target # 2 . miras are generally considered to have the @xmath15 amplitude greater than 0.9 - 1.0 mag ( ita et al . 2004ab ; matsunaga et al . the target # 1 is clearly a mira , whereas the target # 2 falls between miras and semi - regulars . the light curve of the target # 1 shows a clear modulation over the entire observation run indicating a long - term variation which is often observed for carbon - rich miras ( whitelock et al . 2003 ) . we subtract this long - term trend which is fitted by a sine curve with a period of 1500 days . the residual light curve shows a regular periodic variation as expected . then , we applied the phase dispersion minimization ( pdm , stellingwerf 1978 ) to the residual curve to obtain a period 326 days as the best estimate . it should be noted that due to the insufficient coverage of the light curve , especially around the expected minima , we can not exclude the possibility of the star having a period roughly half of the 326 days based only on our @xmath15 band data . nevertheless , the shorter period is unlikely , considering that the star is so red , @xmath18 , and all of such red lmc miras in ita et al . ( 2004 ) have periods of 300@xmath7500 days . we examined various assumptions on the long - term trend ( e.g. , sine curves with different periods and/or amplitudes ) , and the estimated period stays within 15 days from the above value . thus , the period of # 1 is estimated to be @xmath1 days . in contrast , the light curve of the target # 2 appears to show no long - term trend , and the period of 122 days is obtained with the pdm . sine curves with periods of 117@xmath7127 days seems consistent with the photometric data . thus , the periods of # 2 is estimated to be @xmath2 days . for the target # 1 , we obtained near infrared images in the @xmath19 , @xmath20 , and @xmath21-bands using the irsf 1.4-m telescope at south african astronomical observatory ( nagayama et al . our near - infrared observations were carried out on june 14 , 2011 and april 28 , 2012 , and the magnitudes of the target were calibrated based on 2mass magnitudes of the neighboring stars . in addition , the near - infrared magnitudes for our targets were collected from a few large - scale catalogues ( 2mass , skrutskie et al . 2006 ; ukirt dr6 , lawrence et al . 2007 ; denis , epchtein et al . 1994 ) . table [ tab : kband ] lists the magnitudes available from these sources , where the magnitudes were transformed into those in the irsf photometric system ( kato et al . the magnitudes in the denis and ukirt dr6 photometric systems were transformed into the 2mass system by using the equations in carpenter ( 2001 ) and hewett et al . ( 2006 ) , respectively , and then further transformed into the irsf system by using kato et al . the @xmath19 , @xmath20 , and @xmath21 magnitudes list in table [ tab : kband ] and show significant variations . the mean magnitudes are obtained by taking averages of four or three independent data in table 4 . here we check the uncertainty in the transformation between the different catalogs by comparing the magnitudes of stars around our miras after the transformation into the irsf magnitude . the infrared magnitudes well match between the 2mass , ukirt , and irsf photometry measurements within their uncertainties , whereas the @xmath21 magnitudes obtained from the denis are different from those in the others by @xmath22 . we added a shift to the above denis magnitudes to fit them onto the 2mass magnitude scale ( the magnitudes in table 4 are on this corrected scale ) . we estimate the uncertainties in the random - phased mean magnitudes in the @xmath21-band as follows . carbon - rich miras with periods of 300@xmath7350 days in the solar neighborhood and nearby dwarf galaxies appear to have @xmath21-band amplitudes of 0.51.0 mag ( whitelock et al . o - rich miras with periods of 100@xmath7150 days have the @xmath21-band amplitude of [email protected] mag in the sample of whitelock et al . the amplitudes of # 1 and # 2 are assumed to be 0.75 mag and 0.51 mag in the @xmath23-band , respectively , suggesting that the observed mean magnitudes are located within @xmath24 0.38 mag and @xmath24 0.26 mag around their true mean magnitudes , respectively . our infrared datasets consist of four and three independent data for # 1 and # 2 , and thus the uncertainties in the mean magnitudes are approximately 0.19 mag and 0.15 mag . to discuss whether or not our target stars are associated with the sextans dsph , we estimate the distances on the basis of the period - luminosity ( pl ) relation of miras . miras , as well as semi - regulars with a relatively regular variation and an amplitude close to 1 mag , like # 2 , are generally found to lie on the mira pl relation . thus , we can apply this relation to both of our targets ( ita et al . 2004ab ) . adopting the distance modulus of the lmc to be @xmath25 ( alves 2004 ) , the pl relation for the miras ( ita & matsunaga 2011 ) is expressed as @xmath26 ita & matsunaga ( 2011 ) also found that the redder miras tend to show large offsets from the pl relation . they suggested that such offsets are caused by the dust shells around red miras ( @xmath27 ) , and we discuss the implication for our miras in section 3 . for # 1 , the offset in the @xmath21 band is estimated to be @xmath28 mag ( ita & matsunaga 2011 ) . in contrast , the target # 2 is so blue that such an offset is negligible , if any . the interstellar extinctions in the galaxy in the @xmath21 band are calculated to be @xmath29 mag from the map in schlegel et al . ( 1998 ) . after correcting the above offset and the interstellar extinction , we obtain distances of @xmath3 kpc and @xmath4 kpc for the targets # 1 and # 2 . the uncertainties in the distances include those in periods , period - luminosity relation , and mean @xmath21-band magnitudes . the estimated distances are consistent with that of the sextans dsph , @xmath5 kpc ( lee et al . 2003 , 2009 ) . here we discuss the effect of the long - term trend on the distance esitmate for # 1 . long - term trends have been detected from optical and infrared wavelengths , however the relationship among different wavelengths has not been established , in particular when our data points are poor . in addition , the observation dates for the 2mass and the ukirt were mjd 51174 and 53726 , respectively , whereas our @xmath15 observations started on mjd=54836 . the long interval between the infrared and the @xmath15 observations makes the estimation of the long - term trend difficult . however , the long - term trend is suggested to be partly related with absorption by the circumstellar dust ( winters et al . the circumstellar extinction is also suggested to be correlated with the deviation from the pl relation ( ita & matsunaga 2011 ) , and its deviation is corrected using the ( @xmath30 ) color in this work . thus , the above correction considering the ( @xmath30 ) color works , at least partly , as the correction of the long - term trend . in fact , the ukirt photometry indicates that the # 1 was both fainter and redder at its epoch than at the epochs of other photometry . even if we exclude the ukirt magnitudes from our calculation , with the proper correction applied , we get the same result within the uncertainty . we check the reliablity of the distance for # 2 . the ic - band pl relation in ita & matsunaga ( 2011 ) is relatively tight near the period of our mira # 2 ( p=122 days ) , although it has a higher dispersion than the ks - band pl relation . the distance based on the ic - band pl relation is estimated to be @xmath31 kpc , consistent with the distance by the ks - band pl relation ( @xmath4 kpc ) . the systematic radial velocity of the sextans dsph is 226.0 @xmath32 with a dispersion of 8.4 @xmath32 ( battaglia et al . 2011 ) . for the targets # 1 and # 2 , the radial velocities at a single epoch are @xmath33 km s@xmath34 ( mauron et al . 2004 ) and @xmath35 km s@xmath34 ( suntzeff et al . 1993 ) . the radial velocity of a mira measured with the optical spectra shows a time variation due to the pulsation by 1025 @xmath32 ( alvarez et al . the radial velocities of foreground stars in the direction of the sextans dsph show a peak around 050 @xmath32 and a weak tail toward a large velocity . probabilities of foreground stars with the velocities of our targets are low , and this clearly support the memberships to the sextans dsph . another possible system that the targets might be associated with is a stellar stream in the galactic halo . the sdss reveals the presence of a 60-degrees - long stream extending from the ursa major to the sextans ( e.g. , grillmair 2006 ) . however , its heliocentric distance is about 20 kpc , much smaller than those of our targets . the model of law et al . ( 2005 ) predicts the presence of the sagittarius stream at a heliocentric distance of 20 - 60 kpc toward the sextans dsph in the oblate dark - matter halo of the galaxy , although the stream has not been identified in its direction . furthermore , the predicted radial velocities are 250290 km s@xmath34 , which are larger than the measured values of the targets . thus , both of our targets are most likely associated with the sextans dsph . as mentioned in section 1 , our targets in the sextans dsph are expected to be metal - poor . further constraints can be inferred from their locations in the galaxy . the target # 1 is located within the tidal elliptical radius of the sextans dsph ( 160 arcmin , ellipticity of 0.35 ) , but in its outer part . unfortunately , the metallicities for a large sample of stars around the target # 1 are not yet available . nevertheless , battaglia et al . ( 2011 ) reported that the sextans dsph has a clear metallicity gradient : most stars beyond an elliptical radius of 48 arcmin ( ellipticity 0.35 ) in projection appear to have [ fe / h]@xmath362.2 , while metal - rich stars are concentrated toward the center . the target # 1 is located beyond the elliptical radius of 48 arcmin , suggesting that its metallicity is very low . as shown in section 2.4 , the target # 1 is very red , in particular at the ukirt epoch , @xmath37 . ita & matsunaga ( 2011 ) suggested the existence of the dust shells around such red miras from their magnitudes fainter than the pl relation . the deviation of the @xmath21 magnitude in each catalog from that of the period - luminosity relation appears to follow a function of the ( @xmath38 ) and ( @xmath30 ) in ita & matsunaga ( 2011).sloan et al . ( 2009 , 2012 ) and matsuura et al . ( 2007 ) detected the sic dust excess for the spectra of the carbon - rich agb stars in nearby metal - poor dsphs . the color and period of target # 1 are similar to those in the carbon - rich miras in sloan et al . thus , we suggest a possible presence of the circumstellar dust around carbon - rich miras with the metallicity @xmath39 100 times lower than the solar . the sextans dsph has a lower mean metallicity than any other dsph in sloan et al . ( 2012 ) , and the target # 1 can impose an important limit on the metallicity dependence of the dust content around carbon - rich miras . the target # 2 is located in the inner part of the sextans dsph . in contrast to the outer part , relatively metal - rich stars ( [ fe / h]@xmath40 ) coexist with the metal - poor stars at the inner part , and thus the limit on the metallicity of # 2 is not so strong . the short - period and oxygen - rich chemistry suggest that the target # 2 is similar to low - mass miras found in galactic globular clusters . we discovered two miras ( @xmath41 and 122 days , respectively ) toward the metal - poor sextans dsph by performing photometric monitoring for red stars over periods of 3 - 4 years , although the shorter - period one may be a semi - regular . the distances and radial velocities of the objects are consistent with those of the sextans dsph , which suggests that the two objects belong to the sextans dsph . thus , these objects are found in the lowest metallicity dsph with miras , [ fe / h]@xmath42 . the red near - infrared color of the longer - period mira # 1 suggests the presence of circumstellar dust in a system with the lowest metallicity . follow - up spectroscopic observations in the infrared wavelengths is useful to investigate dust formation in the very metal - poor objects . we thank aoki , t. , soyano , t. , tarusawa , k. , and dr . mito , h. for supporting our observations at kiso observatory . we are grateful to dr . feast , m. w. for reading the manuscript carefully and for his comments . ts thanks dr . ita , y. for providing the source code of the period determination . nm acknowledges the support of grant - in - aid for young scientists ( no 23684005 ) from the japan society for the promotion of science ( jsps ) . funding for the sdss and sdss - ii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , the u.s . department of energy , the national aeronautics and space administration , the japanese monbukagakusho , the max planck society , and the higher education funding council for england . the sdss web site is http://www.sdss.org/. adelman - mccarthy j. et al . 2008 , , 175 , 297 alvarez , r. , jorissen , a. , gillet , d. , fokin , a. , & dedecker , m. 2001 , , 379 , 305 alves d. r. 2004 , new astronomy reviews , 48 , 659 battaglia , g. , tolstoy , e. , helmi , a. , irwin , m. , parisi , p. , hill , v. , & jablonka , p. 2011 , , 411 , 1013 carpenter 2001 , , 121 , 2851 cruz k. l. , reid , i. n. , kirkpatrick , j. b. , et al . 2007 , , 133 , 439 epchtein , n. , de batz , b. , copet , e. , et al . 1994 , , 217 , 3 feast , m w. , whitelock , p. a. , & menzies , j. w. 2002 , , 329 , l7 fraser , o. j. , hawley , s. l. , & cook , k. h. 2008 , , 136 , 1242 frogel , j. a. & whitelock , p. a. 1998 , , 116 , 754 grillmair , c. j. 2006 , , 645 , l37 groenewegen , m. a. t. 2000 , , 363 , 901 groenewegen , m. a. t. , sloan , g. c. , soszyski , i. & petersen , e. a. 2009 , , 506 , 1277 habing , h. j. 1996 , , 7 , 97 hewett , p. c. , warren , s. j. , leggett , s. k. , & hodgkin , s. t. 2006 , , 367 , 454 ita , y. , tanabe , t. , matsunaga , n. , et al . 2004a , , 353 , 705 ita , y. , tanabe , t. , matsunaga , n. , et al . 2004b , , 353 , 705 ita , y. & matsunaga , n. 2011 , , 412 , 2345 itoh , n. , soyano , t. , tarusawa , k. , aoki , t. , yoshida , s. , hasegawa , t. , yadomaru , y. , nakada , y. , & miyazaki , s. 2001 , publications of the national astronomical observatory of japan , 6 , 41 jordi , k. , grebel , e. k. , & ammon , k. 2006 , , 460 , 339 kato , d. , nagashima , c. , nagayama , t. , et al . 2007 , , 59 , 615 kirby , e. n. , guhathakurta , p. bolte , m. , sneden , c. , & ghea , m. c. 2009 , , 705 , 328 lagadec , e. , zijlstra , a. a. , sloan , g. c. , et al . 2009 , , 396 , 598 lawrence , a. , warren , s. j. , almaini , o. , et al . 2007 , , 379 , 1599 law , d. r. , johnston , k. v. , & majewski , s. r. 2005 , , 619 , 807 lee , m. g. , park , h. s. , park , j. , et al . 2003 , , 126 , 2840 lee , m. g. , yuk , i. , park , h. s. , harris , j. , & zaritsky , d. 2009 , , 703 , 692 mateo , m. fischer , p. & krzeminski , w. 1995 , , 110 , 2166 matsunaga , m. , fukushi , h. , & nakada y. 2005 , , 364 , 117 matsuura , m. , zijlstra , a. a. , bernard - salas , j. , et al . 2007 , , 382 , 1889 matsuura , m. , barlow , m. j. , zijlstra , a. a. , et al . 2009 , , 396 , 918 mauron , n. , azzopardi , m. , gigoyan , k. , & kendall , t. r. 2004 , , 418 , 77 menzies , j. w. , whitelock , p. a. , feast , m. w. , & matsunaga , n. 2010 , , 406 , 86 menzies , j. w. , feast , m. w. , whitelock , p. a. , & matsunaga , n. 2011 , , 414 , 3492 nagayama , t. et al . 2003 , proc . spie , 4841 , 459 norris , j. e. , wyse , r. f. g , gilmore , g. , yong , d. , frebel , a. , wilkinson , m. i. , belokurov , v. , & zucker , d. b. 2010 , , 723 , 1632 rengstorf , a. w. , thompson , d. l. , mufson , s. l. , et al . 2009 , , 181 , 129 schlegel , d. j. , finkbeiner , d. p. , & davis , m. 1998 , , 500 , 525 skrutskie , m. f. , cutri , r. m. , stiening , r. , et al . 2006 , , 131 , 1163 sloan , g. c. , matsuura , m. , zijlstra , a. a. , lagadec , e. , groenewegen , m. a. t. , wood , p. r. , szyszka , c. , bernard - salas , j. , & van loon , j. th . 2009 , science , 323 , 353 sloan , g. c. , matsuura , m. , lagadec , e. , et al . 2012 , astroph/1204.5754 soszyski , i. , udalsky , a. , szymaski , m. k , et al . 2009 , acta astron . , 59 , 239 stellingwerf , b. f. 1978 , , 224 , 953 suntzeff , n. b. , mateo , m. , terndrup , d. m. , olszewski , e. w. , geisler , d. , & weller , w. 1993 , , 418 , 208 whitelock , p. a. , marang , f. , & feast , m. w. 2000 , , 319 , 728 whitelock , p. a. , feast , m. w. , van loon , j. th . & zijlstra , a. a. 2003 , , 342 , 86 whitelock , p. a. menzies , j. w. , feast , m. w. , matsunaga , n. , tanabe , t. , & ita , y. 2009 , , 394 , 795 winters , j. m. , fleischer , a. j. , gauger , a. , sedlmayr , e. 1994 , , 290 , 623 llllllllll # 1 ( sdss [email protected] ) & 10:15:25.93&@xmath702:04:31.8&@xmath1 & 17.0&3.72 & @xmath3 & @xmath33 & carbon - rich + # 2 ( sdss [email protected])&10:12:34.29&@xmath701:34:40.8&@xmath2 & 15.7&0.94 & @xmath4 & @xmath43&oxygen - rich + sextans dsph & 10:13:03&@xmath701:36.9&&&&@xmath44&@xmath45& + cccccc + mjd & @xmath15 ( mag ) & @xmath46 ( mag ) & mjd & @xmath15 ( mag ) & @xmath46 ( mag ) + 54836.77&17.30&0.08&55646.49&15.53&0.01 + 54875.69&17.20&0.06&55662.53&15.59&0.02 + 54904.57&17.04&0.06&55663.49&15.64&0.03 + 55174.78&18.50&0.13&55694.50&16.08&0.06 + 55221.64&17.99&0.09&55694.50&16.03&0.06 + 55249.61&17.24&0.05&55718.48&16.10&0.06 + 55288.48&16.81&0.03&55879.84&15.03&0.02 + 55314.58&16.87&0.09&55880.82&15.10&0.02 + 55342.49&17.31&0.10&55899.85&14.78&0.01 + 55580.79&15.66&0.07&55899.85&14.78&0.02 + 55580.79&15.65&0.08&55942.73&14.79&0.03 + 55611.56&15.38&0.03&55970.55&14.84&0.03 + ccccccccc + mjd & @xmath15 ( mag ) & @xmath46 ( mag ) & mjd & @xmath15 ( mag ) & @xmath46 ( mag ) & mjd & @xmath15 ( mag ) & @xmath46 ( mag ) + 55249.60&15.54&0.01&55663.49&15.57&0.03&55898.73&15.24&0.04 + 55289.42&15.27&0.03&55694.50&16.11&0.07&55898.73&15.30&0.03 + 55316.47&15.71&0.02&55718.47&15.95&0.08&55939.77&15.81&0.12 + 55344.47&15.87&0.03&55876.77&15.24&0.02&55940.76&15.98&0.05 + 55580.78&15.68&0.08&55877.78&15.18&0.03&55942.72&15.82&0.03 + 55580.78&15.78&0.09&55879.83&15.17&0.05&55967.70&15.70&0.12 + 55611.55&15.55&0.03&55879.84&15.23&0.03&55967.70&15.76&0.10 + 55611.55&15.51&0.05&55880.81&15.24&0.05&55967.70&15.70&0.06 + 55646.65&15.33&0.01&55880.81&15.27&0.04 & & & + 55662.53&15.52&0.03&55898.72&15.28&0.04 & & & + .@xmath19 , @xmath20 , and @xmath21 magnitudes of our targets available in our photometry with the irsf / sirius and other near - infrared catalogs.the magnitudes in other photometric systems were transformed into the irsf system . [ tab : kband ] [ cols="^,^,^,^,^,^,^,^,^ " , ]	we report the discovery of two mira variable stars ( miras ) toward the sextans dwarf spheroidal ( dsph ) galaxy . we performed optical long - term monitoring observations for two red stars in the sextans dsph . the light curves of both stars in the @xmath0 band show large - amplitude ( 3.7 and 0.9 mag ) and long - period ( @xmath1 and @xmath2 days ) variations , suggesting that they are miras . we combine our own infrared data with previously published data to estimate the mean infrared magnitudes . the distances obtained from the period - luminosity relation of the miras ( @xmath3 and @xmath4 kpc , respectively ) , together with the radial velocities available , support memberships of the sextans dsph ( @xmath5 kpc ) . these are the first miras found in a stellar system with a metallicity as low as @xmath6\sim -1.9}$ ] , than any other known system with miras .
one of the most striking features observed at bnl - rhic is the suppression of high @xmath1 production in central a+a collisions relative to peripheral a+a or p+a or p+p collisions . this is generally interpreted as a sign of parton energy loss in hot and dense strongly interacting matter created at the early stage of nucleus - nucleus collisions . this interpretation implies that the suppression should decrease towards lower energies where the initial energy and parton density is expected to be much smaller . numerous results on the energy dependence of hadron yields and spectra indicate that the onset of deconfinement is located at lower sps energies ( see e.g. @xcite ) . existing data on central and peripheral pb+pb collision at @xmath2 from the na49 experiment allow to measure the ratio @xmath6 up to about @xmath7 in transverse momentum ( see @xcite ) . a slight suppression is seen in this range , but the interpretation of this result is hindered by the poorly known interference with the cronin effect ( see @xcite ) . therefore , the nuclear modification factors @xmath3 and @xmath5 would give a clearer picture : the first one is expected to contain a certain amount of cronin effect and a possible suppression , while in the second quantity the cronin effect approximately cancels . therefore , our experiment extracted the modification factors @xmath8 , @xmath9 and @xmath10 from the existing p+p ( see also @xcite ) , p+pb and pb+pb data at top ion - sps energy . the modification factor @xmath10 ( which does not contain the cronin effect to first order ) confirms our previous observations with @xmath6 . our experiment , however , has only limited statistics on p+p and p+pb collision data at top ion - sps energy . therefore , future data runs are planned . the proposed fair - cbm high luminosity experiment has also a great potential in this field , as it will be able to populate the momentum space up to the kinematic limit . for the data analysis strategy of p+p and p+pb collisions , the method developed earlier for pb+pb collisions was adopted ( see @xcite ) . this guarantees a good cancellation of the already small systematic biases . the study was performed in the rapidity interval @xmath11 ( as in pb+pb ) . the only limitation in our data analysis is our relatively poor statistics on p+p and p+pb compared to pb+pb . this also affects our particle identification procedure . however , the area conservation property of the employed poisson maximum - likelihood method ( see @xcite ) for the inclusive disentangling of the energy loss spectra allows to make a good quality @xmath12 analysis even in the low statistics range of momentum space ( see fig . [ dedx ] ) . for completeness , we show the currently available @xmath0 statistics in p+p , p+pb and pb+pb(0 - 5% ) collisions at @xmath2 , and the lack of existing p+p reference data in the nearby energies in fig . ( see also @xcite . ) fig . [ edep ] shows the energy dependence of the nuclear modification factors @xmath3 , @xmath4 and @xmath5 from top ion - sps to top rhic energies ( see @xcite ) . strong increase of the modification factors @xmath3 , @xmath4 is observed at top sps energy . for @xmath5 ( here the cronin effect approximately cancels ) a slight high @xmath1 suppression is observed with binary collision scaling . when using wounded nucleon scaling , the @xmath5 shows an approximate energy - independence at low @xmath1 . both observations confirm our previous similar observations on @xmath6 ( see @xcite ) . however , we are not able to answer the interesting question of higher @xmath1 behavior ( above about @xmath13 ) with the currently existing p+p and p+pb data . as the current p+p and p+pb statistics limits us to @xmath14 , one of our future plans is to record more p+p and p+pb reference data in the na49-future experiment which is in preparation at the cern - sps to increase this range up to 4 gev / c . due to the limitation of this upgraded experiment and the development possibilities of the detector there is no hope in the foreseeable future to reach higher luminosity than 1000 interactions per second at cern . there is however an unique opportunity at the fair accelerator in gsi , darmstadt which will be ready around 2013 , where the planned cbm detector @xcite will be able to tolerate an interaction rate up to @xmath15 which represents an improvement by a factor of about a million relative to the available sps setup . though the proton beam energy at the fair accelerator will be limited to 90 gev this will not decrease the physics interest , because the main issues in p+p and p+a processes are even more exciting at high @xmath1 at this energy , due to the fact that there is little information on either the cronin effect or high @xmath1 suppression in this range . as seen from table [ cbmtable ] , one can profit from the increased luminosity despite the smallness of the estimated cross - sections . even with 45 gev beam energy one expects significant statistics above 4 gev / c . lllll + events & energy [ gev ] & @xmath16 & @xmath17 & @xmath18 + @xmath19 & @xmath20 & @xmath21 & @xmath22 & @xmath23 + + @xmath24 & @xmath20 & @xmath25 & @xmath26 & @xmath27 + + suppression & @xmath28 & @xmath29 & @xmath30 & @xmath31 + @xmath24 & @xmath32 & @xmath33 & @xmath27 & @xmath34 + @xmath35 & @xmath32 & @xmath36 & @xmath37 & @xmath38 + suppression & @xmath39 & @xmath31 & @xmath40 & @xmath41 + @xmath35 & @xmath42 & @xmath37 & @xmath43 & @xmath44 + referring to the `` white region '' in fig . [ stat ] between the lines of 24 and 100 gev beam energy , the cbm experiment will study uncharted territory where the interesting change of the @xmath1 spectra from convex to concave is happening . according to the usual argument this phenomenon is regarded as a simple phase - space limitation effect , but one should be aware of the fact that at 90 gev beam energy the 3.5 gev / c transverse momentum represents only about 0.5 in @xmath45 which is still rather far away from the kinematic boundary . there may be some deeper physical effect , because the deviation from the simple exponential spectrum already starts around 1 gev / c . this was a big surprise at cern - isr in the beginning of the seventies and due to lack of systematic measurements the real cause of this effect was not elucidated since then . of course , this previously unimaginable luminosity which is expected at fair - cbm requires very special detector , data acquisition and trigger systems . the heart of the high @xmath1 filtering system is the sts sub - detector . the basic idea of the filtering algorithm is that one can define rather narrow corridors for the high @xmath1 tracks because they are almost straight . due to the relatively small number of hits in these narrow corridors one can perform extremely fast and exhaustive searches even in case of 1000-fold pileup . in order to ensure full efficiency the corresponding regions of a given silicon surface ( = mosaic ) can be used in a number of corridors , therefore a so - called `` multi - tasking mosaic - trigger network '' is proposed which is schematically shown in fig . [ cbmfigure ] . first na49 results on the nuclear modification factors @xmath3 , @xmath4 and @xmath5 were presented for the particle species @xmath0 at @xmath2 beam energy , based on a study on p+p , p+pb and pb+pb collisions . the current statistical limitation of our p+p and p+pb dataset is pointed out , and future plans are discussed . the experiment fair - cbm is introduced as a possible continuation of this low energy p+p , p+a , a+a programme at very high @xmath1 . development work is in progress in the framework of the future - daq project as part of the eu fp6 contract nr . 506078 ( hadronphysics ) .	transverse momentum spectra of @xmath0 at midrapidity are measured at high @xmath1 in p+p and p+pb collisions at @xmath2 beam energy by the na49 experiment . this study is complementary to our previous results on the same spectra from pb+pb collisions . the nuclear modification factors @xmath3 , @xmath4 and @xmath5 as a function of @xmath1 are extracted and compared to rhic measurements , thus providing insight into the energy dependence of nuclear modification . the modification factor @xmath5 proved to be consistent with our previous results on the central to peripheral modification factor @xmath6 . the limitation of our current @xmath1 range is discussed and planned future upgrades are outlined . some aspects of the fair - cbm experiment are also presented as a natural future continuation of the measurements at very high @xmath1 .
t hooft @xcite and mandelstam @xcite proposed long ago that quark confinement in qcd would come about as the result of the confinement of color electric flux into flux tubes , and that such flux tubes would form in a dual superconductor . one possible pathway to formation of this dual superconductor was offered later by t hooft @xcite . he showed that an abelian projection of a non - abelian gauge theory contains magnetic monopoles . though the effective interaction among these monopoles is hard to calculate , it is not unreasonable to suppose that they form a condensate like that of the cooper pairs in a superconductor . magnetic _ condensate would then bring about an _ electric _ meissner effect , confining electric flux and electric charge . in order to study particle dynamics in qcd , one has to study dynamics of the flux tube . casher , kogut , and susskind @xcite argued that particle creation in the flux tube is the soft process responsible for the inside outside cascade in @xmath0 annihilation ; casher , neuberger , and nussinov @xcite calculated the particle spectrum via wkb ( generalizing schwinger s @xcite famous formula ) , and this picture then entered phenomenology via the lund monte carlo program @xcite and its descendants . the flux tube has also been much studied in the context of @xmath1 @xcite and @xmath2 @xcite collisions . these and subsequent studies generally lacked any dynamics for the structure of the flux tube itself . we have taken the first step of studying the dynamics of classical charges moving in an electric flux tube and the reaction of the flux tube in the dual superconductor @xcite . to specify the model , we begin with maxwell s equations coupled to both magnetic and electric currents , @xmath3 eq . ( [ mgauss4 ] ) is no longer just a bianchi identity ; thus a vector potential can be introduced only if a new term is added to take care of the magnetic current , _ viz . , _ @xmath4 this vector potential can be coupled to electric charges as usual ; in order to introduce _ magnetic _ charges , one introduces a dual potential @xcite via @xmath5 now we can write a model for the monopoles , for which the simplest is an abelian higgs theory @xcite , @xmath6 where @xmath7 this theory should produce the desired magnetic condensate to confine electric charge . the magnetic current appearing in ( [ mgauss4 ] ) is @xmath8 we envision a long flux tube with some charge distribution @xmath9 at the ends . far from the ends , the flux tube is initially the well - known cylindrically symmetric solution of the field equations above ( and we impose cylindrical symmetry on the subsequent evolution ) . we release into this flux tube a fluid of electrically charged particles , realized via simple two - fluid mhd .. ] the two fluids ( @xmath10 and @xmath11 ) obey the euler equations @xmath12 = \pm e \be \pm e \bv^{\pm } \times \bh - \frac{1}{n_e}{\nabla p}\ ] ] and the continuity equation @xmath13 the electric current is thus @xmath14 this fluid will flow under the influence of the initial electric field and will screen the charges at the ends ; the fluid s inertia will set up plasma oscillations . since the flux tube geometry is dynamic , the tube itself will react to the weakening of the field by contracting under the pressure of the vacuum field @xmath15 ; then it will be forced open as the fluid overshoots and oscillates . the final ingredient is the equation of state of the fluid , which we take to be that of a quark gluon plasma . the superconductor has two length scales , the vector mass @xmath16 and the scalar ( higgs ) mass @xmath17 . in condensed matter language , we have the london penetration depth @xmath18 and the coherence length @xmath19 ; their ratio @xmath20 determines whether we have a type i ( @xmath21 ) or a type ii ( @xmath22 ) superconductor . another scale in our problem , introduced by the charged fluid , is the plasma frequency @xmath23 . if @xmath24 , one expects that electromagnetic radiation will be unable to penetrate into the superconductor , with the meissner effect providing dynamical confinement as well as static ; the regime @xmath25 , where photons ( i.e. , abelian gluons ) propagate freely is presumably outside the range of applicability of the model . our numerical results , presented in the figures , belie our expectations . we show in figure 1 the on - axis electric field for the type i case , with @xmath24 . there is strong non - linear modification of the plasma oscillations . figure 2 presents snapshots of the electric field @xmath26 and the higgs field @xmath27 . it is clear that electric flux penetrates far outside the initial radius of the flux tube , accompanied by strong oscillations in @xmath28 . ( the type ii case is not too different , though less spectacular . ) this surely casts doubt on the usefulness of this model for the study of dynamical confinement phenomena . 9 g. t hooft , nucl . b190 ( 1981 ) 455 . s. mandelstam , phys . 23c ( 1973 ) 245 . g. t hooft , nucl . b190 ( 1981 ) 455 . a. casher , j. kogut and l. susskind , phys . rev . d 10 ( 1974 ) 732 . a. casher , h. neuberger , and s. nussinov , phys . d 20 ( 1979 ) 179 ; _ ibid . _ 21 ( 1980 ) 1966 . j. schwinger , phys . 82 ( 1951 ) 664 . b. andersson , g. gustafson , g. ingelman , and t. sjostrand , phys . 97 ( 1983 ) 31 ; b. andersson , the lund model , cambridge university press , cambridge , 1998 . a. k. kerman , t. matsui , and b. svetitsky , phys . 56 ( 1986 ) 219 . t. s. bir , h. b. nielsen , and j. knoll , nucl . phys . b245 ( 1984 ) 449 . f. cooper , j. m. eisenberg , y. kluger , e. mottola , and b. svetitsky , phys . d * 48 * , 190 ( 1993 ) ; j. m. eisenberg , phys . d * 51 * , 1938 ( 1995 ) . m. a. lampert and b. svetitsky , hep - ph/9905455 . d. zwanziger , phys . d 3 ( 1971 ) 880 ; m. blagojevi and p. senjanovi , phys . 157 ( 1988 ) 233 . t. suzuki , prog . theor . phys . 80 ( 1988 ) 929 ; s. maedan and t. suzuki , _ ibid . _ 81 ( 1989 ) 229 ; t. suzuki , _ ibid . _ 81 ( 1989 ) 752 . h. suganuma , s. sasaki , and h. toki , nucl . b435 ( 1995 ) 207 .	we have studied plasma oscillations in a flux tube created in a dual superconductor . starting from a static flux tube configuration , with electric charges at either end , we release a fluid of electric charges in the system that accelerate and screen the electric field . the weakening of the electric field allows the flux tube to collapse , and the inertia of the charges forces it open again . we find strong radiation of electric flux into the superconductor in all regimes of couplings .
black holes are important objects in physics especially for quantum gravity , they have the analogy of laws between dynamics and thermodynamics ( @xcite ) . the temperature and entropy are analogous to the surface gravity and area of black hole . the four laws of thermodynamics in black hole have been discussed ( ) , and generalized in ads spacetime ( @xcite ) , kerr - ads black hole , and kerr - newman - ads black holes . the spin entropy of rotating black hole have also been discussed ( @xcite ) . if the black hole exists cauchy horizon and event horizon , we could study the products of these quantities , such as horizon and surface gravity etc . the products of the thermodynamical features for black hole with rotational symmetry have been discussed ( @xcite ) . it is shown that sometimes the products of black hole are independent with adm ( arnowitt - deser - misner ) mass , but depend on charge and spin . through studying these products , we could know these may turn to universal and hold for other solution for nontrivial surrounding black holes . the phase transition in black hole has an interesting topics ( ) . for the schwarzschild black hole the heat capacity is always negative and no phase transition exists . but for rn black hole , the phase transition will happen due to the existence of charge . if black hole has angular momentum and charge , the second order phase transition will happen . if black hole is surrounded by quintessence or other matters , the phase transition maybe happen . for the quintessence matter ( other energy matters ) around schwarzschild black hole , the einstein equation s solution have been obtained by kiselev ( @xcite ) . these solution are determined by the parameters @xmath0 and @xmath1 . interestingly these solutions could also describe radiation or dust around black holes ( @xcite ) . following these works , the particle collision have been discussed ( @xcite ) . the thermodynamical relation and phase transition have been obtained in kiselev black hole for different situations ( @xcite ) . recently the rotation kiselev solution and kerr - newman kiselev solution have been obtained ( ) . in this paper we want to study the thermodynamical relation in rotational kiselev black hole , the products of thermal quantities , the generalized smarr formula and christodoulou - ruffini mass formula . we also calculate the phase transition for these black holes in the case of @xmath4 . the outline of this paper is as follows . in section 2 , we discuss the basic aspects of rotational kiselev black hole by one order perturbation for horizon . the results show that the products are determined by the parameters @xmath0 and @xmath1 . we also generalize the smarr mass formula and christodoulou - ruffini mass formula in subsection 2.1 . in section 3 , the phase transition have been discussed , specially for radiation situation @xmath4 . the summary is presented in the last section . in this paper we set @xmath5 . the rotational symmetry solution for einstein s equations in quintessence ( or other energy matter ) have been obtained by toshmatov ( @xcite ) @xmath6 where @xmath7 in this spacetime metric , @xmath8 and @xmath9 are the mass and spin of the black hole . @xmath1 is the quintessence ( or other matters ) parameter . the state parameter @xmath0 is given by the equation of state @xmath10 . if the energy matter is quintessence , @xmath0 will be @xmath11 to lead the universe acceleration . the case of @xmath4 represents radiation , and the case of @xmath12 represents dust around the kerr black hole . here we will discuss general situation , specially for above three situations . the black hole horizon is determined by the following equation @xmath13 this equation has three horizons if @xmath11 , including the horizon , outer horizon and cosmological horizon . in the case of @xmath14 , the cosmological horizon will vanish and the equation ( [ 3 ] ) exists two horizons . here we only consider the inner and outer horizons . the cosmological horizon does not exist , but this situation is also interesting because its metric describes other energy matters such as radiation and dust around the kerr black hole . because @xmath1 is small , we make the perturbation method for @xmath1 in horizon equation ( @xcite ) . in this paper we make the quantities relating to @xmath1 one - order , the horizon is written as @xmath15 where @xmath16 is the horizon of kerr black hole . @xmath17 is the outer horizon or event horizon @xmath18 and @xmath19 is the inner horizon or canchy horizon @xmath20 . substituting ( [ 4 ] ) into ( [ 3 ] ) , we expand the equation to the first order term . the equation becomes @xmath21 the solution of equation ( [ 5 ] ) satisfies the condition of @xmath22 for @xmath23 , therefore the solution is given by @xmath24 and @xmath25 @xmath26 here we use three special examples with different @xmath0 to show the horizon and product of horizon . the product of horizon for @xmath0 is given by @xmath27 case one : @xmath28 corresponding to quintessence , there is no exact formalism horizon for @xmath29 . form equation ( [ 7 ] ) and ( [ 8 ] ) , we find @xmath30 and the product of the horizon is given by @xmath31 case two : @xmath4 representing to radiation , there are two exact horizon expressions for @xmath29 . from equation ( [ 3 ] ) , we find that @xmath32 and the product of the horizon is given by @xmath33 case three : @xmath12 relating to dust , there is exact formalism for two horizons . form equation ( [ 3 ] ) , we find @xmath34 and the product of the horizon is given by @xmath35 in summary , the product of horizon is determined by @xmath0 and @xmath1 . for the quintessence around kerr black hole , the product of horizon is mass dependent . when @xmath23 , the result is opposite . for the radiation and dust around kerr black hole , the products of horizons are independent on mass , but depend on spin and @xmath1 . for stationary and axial symmetric black hole , the areas of the inner and outer horizons of the black hole are given by @xmath36 the semi - classical bakenstein - hawking entropy at @xmath37 is given by ( @xcite ) @xmath38 the surface gravity of the black hole is defined by @xmath39 , \label{18}\ ] ] where @xmath40 . from the black hole thermodynamics , the hawking temperature for @xmath37 is @xmath41 . \label{19}\ ] ] for rotation black hole , the rotation velocity of the horizon is @xmath42 using @xmath43 and @xmath44 , we can calculate the komar energy ( @xcite ) , the product of surface gravity , the surface temperatures , the areas and entropy at @xmath37 , which are given by the komar energy @xmath45 , \label{21}\ ] ] the product of surface gravity and surface temperature @xmath46 , \label{22}\ ] ] the product of komar energy @xmath47 , \label{23}\ ] ] the product of areas and entropy @xmath48 from the equation ( [ 22 ] ) , ( [ 23 ] ) and ( [ 24 ] ) , we find that the products of surface gravity , the surface temperatures and the komar energy are related to the mass of black hole for different @xmath0 . this property is the same with that of spherically symmetric black hole ( @xmath49 ) . the relation between the products of area ( entropy ) and the mass is determined by @xmath0 and @xmath1 . when @xmath23 , the products of area and entropy are independent on mass . here we give the results for three cases : quintessence , radiation and dust . case one : the quintessence with @xmath28 , the products become @xmath50 @xmath51 @xmath52 case two : the radiation with @xmath4 , the products become @xmath53 @xmath54 @xmath55 case three : the dust with @xmath12 , the products are @xmath56 @xmath57 @xmath58 from the formula ( [ 15 ] ) , the general formalism ( @xcite ) relating to the entropy , angular momentum and mass of the black hole for any @xmath0 is @xmath59 because the equations are very complex for general @xmath0 , we do nt get the mass formula . but for special @xmath0 , we will calculate the mass formula and derive the state equation . here we will calculate the special case : the radiation with @xmath4 around kerr black hole . for simple , we only consider the property on the outer horizon . the event horizon has exact formalism , and the euqation ( [ 44 ] ) becomes @xmath60 we then get the function of @xmath61 as @xmath62 we define the quantities relating to @xmath63 and @xmath1 , the effective surface tension and angular velocity are given by @xmath64 @xmath65 according to the first law of thermodynamics , the mass change of black hole is determined by the areas and angular momentum as @xmath66 where @xmath67 is effective surface tension given by @xmath68 @xmath69 and @xmath70 . the angular volocity @xmath71 is given by @xmath72 the black hole mass can decrease or increase , but the irreducible mass @xmath73 of black hole can not decrease ( @xcite ) . the increase of @xmath73 will lead to many physical processes in black hole . according to hawking , the black hole satisfies the thermodynamics second law . the areas of horizon or entropy all increase , e.g. , @xmath74 , leading a relation between horizon area and irreducible mass for black hole . in general quintessence , the black hole has regular event horizon and cauchy horizons , thus the irreducible mass and horizon areas satisfy the following equation @xmath75 where @xmath76 and @xmath77 are black hole irreducible masses on the outer and inner horizons . the product of the irreducible masses is given by @xmath78 the rest mass of rotating black hole is given by the following expression @xmath79 in this section , we study the phase transition in these black holes . the heat capacity of the black hole is an interesting thermodynamical property presenting the stability and instability of the black hole . general black hole has a negative heat capacity to be unstable and produce hawking radiation ( @xcite ) . but if the black hole brings angular momentum or charge , the heat capacity could be positive , and the phase transition will occur . ( @xcite ) . for the quintessence matter ( other energy matter ) around spherical symmetry black hole , the phase transition has been discussed ( ) . for rotational kiselev black hole or kerr - newman black hole , we want to analyze the phase transition . in order to obtain the equation of state for thermodynamical system . substituting ( [ 15 ] ) into ( [ 3 ] ) , we write the black hole mass as a function of black hole entropy and angular momentum @xmath80 given by @xmath81 from the section 2 , we know that the first law of thermodynamics for these black holes is given by @xmath82 using the relation ( [ 130 ] ) , we define other thermodynamical quantities given by @xmath83 @xmath84 in order to simplify the equation , we set @xmath85 @xmath86 where @xmath87 is given by @xmath88 . \label{136}\ ] ] the phase transition will produce when the following condition satisfies @xmath89 for general @xmath0 , we can not give the expression of @xmath90 . here we take @xmath4 to study the feature of phase transition . the generalized smarr mass function has been obtained in section 2.1 as @xmath91 using this equation , we get the expression for @xmath92 as @xmath93 @xmath94 @xmath95 the phase transition will produce when the following equation satisfies @xmath96 in order to analyze the critical point of the phase transition , we set @xmath97 to obtain @xmath98 and @xmath99 . the equation becomes @xmath100 [ cols="<,^,^,^,^,^,^,^,^",options="header " , ] [ tab : ks ] through numerical calculation , we find that the phase transition with @xmath101 demands @xmath102 . for different @xmath1 , the critical black hole spin @xmath9 is given in table 1 . we find that the existence of other energy matters will boost the critical black hole spin @xmath9 . when @xmath23 it reduces to kerr black hole case and the phase transition is second order phase transition ( @xcite ) . when @xmath49 it reduces to spherically symmetric case and the phase transition is also second order phase transition ( ) . for the rotational black hole in radiation case the phase transition is second order phase transition . versus entropy @xmath103 for different angular momentum ( j=0.2,j=0.5,@xmath4),title="fig : " ] versus entropy @xmath103 for different angular momentum ( j=0.2,j=0.5,@xmath4),title="fig : " ] versus entropy @xmath103 at constant angular momentum for different parameter @xmath1 . ( @xmath4),title="fig : " ] versus entropy @xmath103 at constant angular momentum for different parameter @xmath1 . ( @xmath4),title="fig : " ] we plot the variation of the semi - classical hawking temperature @xmath67 versus the entropy @xmath103 for different parameters @xmath1 and angular momentum @xmath104 in picture 1 . the variation of the heat capacity @xmath90 versus the entropy @xmath103 for different parameters @xmath1 is shown in picture 2 . from picture 1 , we find that when @xmath1 decreases the semi - classical hawking temperature will also decrease . from picture 2 , we find that the heat capacity will pass a critical point with the entropy , representing the phase transition point . interestingly , the critical point shifts to higher entropy when the parameter @xmath1 increases . we have discussed some thermodynamical features on two horizons for rotation kiselev black hole . by horizon perturbation , we obtain black hole area , surface gravity , entropy and so on . through calculating the products of these quantities , we find that they depend on the parameters @xmath0 and @xmath1 , and are different from ones in kerr black hole . we also generalize the smarr mass formula and study the christodoulou - ruffini mass formula , and find that the first law of thermodynamics is also satisfied . finally we calculate the phase transition for the black hole with @xmath4 , and find that the phase transition point shifts to higher entropy when the parameter @xmath1 increases .	we calculate the thermodynamical features of rotational kiselev black holes , specifically we use one order approximate of horizon to calculate thermodynamical features for all @xmath0 . the thermodynamics features include areas , entropies , horizon radii , surface gravities , surface temperatures , komar energies and irreducible masses at the cauchy horizon and event horizon . at the same time the products of these features have been discussed . we find that the products are independent with mass of black hole and determined by @xmath0 and @xmath1 . the features in the situations of @xmath2 and @xmath3 ( quintessence matter , radiation and dust ) have been discussed in detail . we also generalize the smarr mass formula and christodoulou - ruffini mass formula to these black holes . finally we study the phase transition for black holes with different @xmath0 and obtain the state equation . we analyze the phase transition for @xmath4 , and find that @xmath1 shifts the critical point of phase transition .
the actual measurement performed by the e821 collaboration is of the muon s _ anomalous _ magnetic moment , which is to say , the coefficient @xmath0 of the non - renormalizable operator @xmath8 within the standard model , @xmath0 is predicted to be @xcite @xmath9 it has been claimed that the quoted uncertainty is actually too small @xcite ; the dominant uncertainty is the hadronic contribution to the photon polarization diagrams which is extracted from experimental measurements of @xmath10 in the vicinity of the low - energy meson resonances . however , strong arguments have been made to reinforce the quoted values @xcite , and we will accept them here as given . the measurement made by e821 is @xcite : @xmath11 from which one deduces a discrepancy between the experiment and the standard model of @xmath12 that is , the measured value is _ larger _ than the prediction by @xmath13 . the susy contributions to @xmath0 have been known since the early days of susy and have become more complete with time@xcite . in this paper we will follow the notation of ref . @xcite which has the advantage of using the standard conventions of haber and kane @xcite ; any convention which we do not define here can be found in either of these two papers . before we begin the discussion of our work , let us review briefly a few of the analyses to date . the first paper to use the new e821 data in the context of susy was that of czarnecki and marciano @xcite who only attempted to approximately bound the sparticle spectrum . more complete analyses followed quickly thereafter by everett @xcite and feng and matchev @xcite . the former @xcite argued that only very large values of @xmath2 were generically consistent with the data ; given those , a @xmath14 upper bound of @xmath15 could be placed on some sparticle ( gaugino or slepton ) for @xmath16 , and @xmath17 for both a chargino and the muon sneutrino . the latter analysis @xcite found a model - independent @xmath1 limit on the mass of the lightest `` observable '' sparticle of @xmath18 for @xmath19 . martin and wells did a more complete analysis , but focussed on the lighter chargino and charged smuon . their model - independent analysis found no upper bound on the chargino mass at @xmath1 but found @xmath20 . with the added assumption of gaugino mass unification ( , @xmath21 ) , the lighter chargino was bounded by about @xmath22 at @xmath23 . the lack of a strong upper bound on the chargino was due to the neutralino contributions at small @xmath2 ( all the way down to @xmath24 ) which can explain the e821 data without any chargino piece at all . finally , there have been a number of other papers @xcite which have studied the susy contributions within specialized scenarios which we will not discuss here . in the mass eigenbasis , there are only two one - loop susy diagrams which contribute to @xmath0 , shown in figure 1 . the first has an internal loop of smuons and neutralinos , the second a loop of sneutrinos and charginos . but the charginos , neutralinos and even the smuons are themselves admixtures of various interaction eigenstates and we can better understand the physics involved by working in terms of these interaction diagrams , of which there are many more than two . we can easily separate the leading and subleading diagrams in the interaction eigenbasis by a few simple observations . = 4.25 in first , the magnetic moment operator is a helicity - flipping interaction . thus any diagram which contributes to @xmath0 must involve a helicity flip somewhere along the fermion current . this automatically divides the diagrams into two classes : those with helicity flips on the external legs and those with flips on an internal line . for those in the first class , the amplitude must scale as @xmath25 ; for those in the second , the amplitudes can scale instead by @xmath26 , where @xmath27 represents the mass of the internal susy fermion ( a chargino or neutralino ) . since @xmath28 , it is the latter class that will typically dominate the susy contribution to @xmath0 . therefore we will restrict further discussion to this latter class of diagrams alone . secondly , the interaction of the neutralinos and charginos with the ( s)muons and sneutrinos occurs either through their higgsino or gaugino components . thus each vertex implies a factor of either @xmath29 ( the muon yukawa coupling ) or @xmath30 ( the weak and/or hypercharge gauge coupling ) . given two vertices , the diagrams therefore scale as @xmath31 , @xmath32 or @xmath33 . in the standard model , @xmath29 is smaller than @xmath30 by roughly @xmath34 . in the minmal susy standard model ( mssm ) at low @xmath2 , this ratio is essentially unchanged , but because @xmath29 scales as @xmath35 , at large @xmath2 ( @xmath36 ) the ratio can be reduced to roughly @xmath37 . thus we can safely drop the @xmath31 contributions from our discussions , but at large @xmath2 we must preserve the @xmath32 pieces as well as the @xmath33 pieces . the pieces that we will keep are therefore shown in figures 2 . in fig . 2(a)-(e ) are shown the five neutralino contributions which scale as @xmath33 or @xmath32 ; in fig . 2(f ) is the only chargino contribution , scaling as @xmath32 . the contributions to @xmath0 from the @xmath38th neutralino and the @xmath39th smuon due to each of these component diagrams are found to be : @xmath40 and for the @xmath41th chargino and the sneutrino : @xmath42 the matrices @xmath43 , @xmath44 and @xmath45 are defined in the appendix along with the functions @xmath46 . a careful comparison to the equations in the appendix will reveal that we have dropped a number of complex conjugations in the above expressions ; it has been shown previously @xcite that the susy contributions to @xmath0 are maximized for real entries in the mass matrices and so we will not retain phases in our discussion . = 4.25 in in many of the previous analyses of the mssm parameter space , it was found that it is the chargino - sneutrino diagram at large @xmath2 that can most easily generate values of @xmath47 large enough to explain the observed discrepancy . from this observation , one can obtain an upper mass bound on the lightest chargino and the muon sneutrino . however , this behavior is not completely generic . for example , martin and wells have emphasized that the @xmath48 neutralino contribution can _ by itself _ be large enough to generate the observed excess in @xmath0 , and since it has no intrinsic @xmath2 dependence , they could explain the data with @xmath2 as low as 3 . we can reproduce their result in a simple way because the @xmath48 contribution has a calculable upper bound at which the smuons mix at @xmath49 , @xmath50 , and @xmath51 with @xmath52 . then @xmath53 where we have used the fact that @xmath54 and @xmath55 and have included a 7% two - loop suppression factor . though any real model will clearly suppress this contribution somewhat , this is still @xmath56 times larger than needed experimentally . this pure @xmath48 scenario is actually an experimental worst - case , particularly for hadron colliders . the only sparticles that are required to be light are a single neutralino ( which is probably @xmath57-like ) and a single @xmath58 . the neutralino is difficult to produce , and if stable , impossible to detect directly . the neutralino could be indirectly observed in the decay of the @xmath58 as missing energy , but production of a @xmath58 at a hardon machine is highly suppressed . in the worst of all possible worlds , e821 could be explained by only these two light sparticles , with the rest of the susy spectrum hiding above a tev . further , even the `` light '' sparticles can be too heavy to produce at a @xmath59 linear collider . while this case is in no way generic , it demonstrates that the e821 data by itself does not provide any sort of no - lose theorem for run ii of the tevatron or even for the lhc and nlc . this raises an important experimental question : how many of the mssm states must be `` light '' in order to explain the e821 data ? in the worst - case , it would appear to be only two . even in the more optimistic scenario in which the chargino diagram dominates @xmath47 , the answer naively appears to be two : a single chargino and a single sneutrino . in this limit , @xmath60 where we have bounded @xmath61 by 10 by assuming @xmath62 . but this discussion is overly simplistic and we can do much better , as we will see . there are a total of 9 separate sparticles which can enter the loops in fig . [ simple ] : 1 sneutrino , 2 smuons , 2 charginos and 4 neutralinos . the mass spectrum of these 9 sparticles is determined entirely by 7 parameters in the mssm : 2 soft slepton masses(@xmath63 , @xmath64 ) , 2 gaugino masses ( @xmath65 , @xmath66 ) , the @xmath67-term , a soft trilinear slepton coupling ( @xmath68 ) and finally @xmath2 . of these , @xmath68 plays almost no role at all and so we leave it out of our discussions ( see the appendix ) . and in some well - motivated susy - breaking scenarios , @xmath65 and @xmath66 are also correlated . thus there are either 5 or 6 parameters responsible for setting 9 sparticle masses . there are clearly non - trivial correlations among the masses which can be exploited in setting mass limits on the sparticles . first , there are well - known correlations between the chargino and neutralino masses ; for example , a light charged @xmath69 implies a light neutral @xmath70 and vice - versa . there are also correlations in mixed systems ( , the neutralinos , charginos and smuons ) between the masses of the eigenstates and the size of their mixings . consider the case of the smuons in particular ; their mass matrix is given in the appendix . on diagonalizing , the left - right smuon mixing angle is given simply by : @xmath71 the chargino contribution is maximized for large smuon mixing and large mixing occurs when the numerator is of order or greater than the denominator ; since the former is suppressed by @xmath25 , one must compensate by having either a very large @xmath67-term in the numerator or nearly equal @xmath72 and @xmath73 in the denominator , both of which have profound impacts on the spectrum . there is one more correlation / constraint that we feel is natural to impose on the mssm spectrum : slepton mass universality . it is well - known that the most general version of the mssm produces huge flavor - changing neutral currents ( fcncs ) unless some external order is placed on the mssm spectrum . by far , the simplest such order is the demand that sparticles with the same gauge quantum numbers be degenerate . thus we expect @xmath74 and @xmath75 . then the mass matrix for the stau sector is identical to that of the smuons with the replacement @xmath76 in the off - diagonal elements . this enhancement of the mixing in the stau sector by @xmath77 implies that @xmath78 . in particular , if @xmath79 then @xmath80 and qed will be broken by a stau vev . given slepton universality , this imposes a constraint on the smuon mass matrix : @xmath81 or on the smuon mixing angle : @xmath82 where @xmath83 are the positive roots of @xmath84 . while not eliminating the possibility of @xmath85 , this formula shows that a fine - tuning of at least 1 part in 17 is needed to obtain @xmath86 mixing . we will not apply any kind of fine - tuning criterion to our analysis , yet we will find that this slepton mass universality constraint sharply reduces the upper bounds on slepton masses which we are able to find in our study of points in mssm parameter space . ( as an aside , if one assumes slepton mass universality at some susy - breaking messenger scale above the weak scale , yukawa - induced corrections will break universality by driving the stau masses down . this effect would further tighten our bounds on smuon masses and mixings . ) the above discussion has an especially large impact on the worst - case scenario in which the @xmath48 contibutions dominates @xmath47 . for generic points in mssm parameter space , one expects that @xmath87 which reduces the size of the @xmath48 contribution by a factor of 17 . as a byproduct , the masses required for explaining the e821 anomaly are pushed back towards the range that can be studied by a @xmath59 linear collider . now that we have established the basic principle of our analysis , let us carry it out in detail . we will concentrate on three basic cases . the first case is the one most often considered in the literature : gaugino mass unification . here one assumes that the weak - scale gaugino mass parameters ( @xmath65 and @xmath66 ) are equal at the same scale at which the gauge couplings unify . this implies that at the weak scale @xmath88 . the second case we consider is identical to the first with the added requirement that the lightest susy sparticle ( lsp ) be a neutralino . this requirement is motivated by the desire to explain astrophysical dark matter by a stable lsp . finally we will also consider the most general case in which all relevant susy parameters are left free independent of each other ; we will refer to this as the `` general mssm '' case . the basic methodology is simple : we put down a logarithmic grid on the space of mssm parameters ( @xmath65 , @xmath66 , @xmath63 , @xmath64 and @xmath67 ) for several choices of @xmath2 . the grid extends from @xmath89 for @xmath65 , @xmath66 and @xmath67 , and from @xmath90 for @xmath91 , up to @xmath92 for all mass parameters . for the case in which gaugino unification is imposed @xmath66 is no longer a free parameter and our grid contains @xmath93 points . for the general mssm case our grid contains @xmath94 points . only @xmath95 is considered since that maximizes the value of @xmath47 . finally , for our limits on @xmath2 we used an adaptive mesh routine which did a better job of maximizing @xmath47 over the space of mssm inputs . by running with grids of varying resolutions and offsets we estimate the error on our mass bounds to be less than @xmath96 . perhaps the most important information that can be garnered from the e821 data is an upper bound on the scale of sparticle masses . in particular , one can place upper bounds on the masses of the lightest sparticle(s ) as a function of @xmath47 . previous analyses have often followed this approach , deriving upper bounds on the lightest sparticle from among the gauginos and sleptons , or even more specifically , from among the charginos and smuons . we too derive bounds on the lightest sparticles , but as we have argued in the previous section , we can also derive bounds on the second , third , and even fourth lightest sparticles through the mass correlations . these bounds on additional light sparticles provide an important lesson . without them there remains the very real possibility that the e821 data is explained by a pair of light sparticles and that the remaining susy spectrum is out of reach experimentally . but our additional bounds will give us some indication not only of where we can find susy , but also of how much information we might be able to extract about the fundamental parameters of susy the more sparticles we detect and measure , the more information we will have for disentangling the soft - breaking sector of the mssm . in fig . [ massfig ] we have shown the upper mass bounds for the lightest four sparticles assuming gaugino mass unification . these bounds are not bounds on individual species of sparticles ( which will come in the next section and always be larger than these bounds ) but simply bounds on whatever sparticle happens to be lightest . the important points to note are : _ ( i ) _ the maximum values of the mass correspond to the largest value of @xmath2 , which is to be expected given dominance of the chargino diagram at large @xmath2 ; _ ( ii ) _ the @xmath1 limit ( central value ) of the e821 data requires _ at least _ 4 sparticles to lie below roughly @xmath97 ; and _ ( iii ) _ for low values of @xmath2 a maximum value of @xmath47 is reached ( we will return to this later ) . the same plots could be produced with the additional assumption that the lsp be a neutralino , but we will only show the case for the lsp bound , in fig . [ lspfig ] . in this figure , the solid lines correspond to a neutralino lsp , while the dotted lines are for the more general case discussed above ( , they match the lines in fig . [ massfig](a ) ) . notice that for @xmath98 there is little difference between the cases with and without a neutralino lsp . furthermore , at the extreme upper and lower values of @xmath2 there is little difference . it is only for the intermediate values of @xmath2 that the mass bound shifts appreciably ; for @xmath99 it comes down by as much as @xmath90 compared to the more general case . = 4.25 in finally , we consider the most general mssm case , , without gaugino unification . here the correlations are much less pronounced , but interesting bounds still exist . for example the central value of the e821 data still demands at least 3 sparticles below @xmath59 ( rather than four for the previous cases ) . in figure [ nongutmaxes ] we demonstrate this explicitly by plotting the masses of the four lightest sparticles for @xmath100 and a wide range of @xmath47 . we see that dropping the gaugino unification requirement has one primary effect : the mass of the lsp is significantly increased . this is because the lsp in the unified case is usually a @xmath101 but is nt itself responsible for generating @xmath47 . in the general case , the lsp must participate in @xmath47 ( otherwise its mass could be arbitrarily large ) and so is roughly the mass of the _ second _ lightest sparticle in the unified case , whether that be a @xmath58 or @xmath102 . otherwise the differences between the more general mssm and the gaugino unified mssm are small . in particular we still find that at least 4 sparticles must be light , though the @xmath1 bound of @xmath22 for the unified case extends now slightly to @xmath3 . = 4.25 in we have summarized all this data on the lsp in table [ masstable ] where we have shown the mass bounds ( using both the @xmath1 limit and the central value of the e821 data ) on the lsp for various @xmath2 values and with our various assumptions . the last line in the table represents an upper bound for any model with @xmath103 : @xmath104 for the e821 @xmath1 lower bound ( central value ) of @xmath47 . but perhaps of equal importance are the bounds on the next 3 lightest sparticles ( the `` 2lsp , '' `` 3lsp '' and `` 4lsp '' ) : @xmath105 , @xmath106 and @xmath107 . thus there must be at least 4 sparticles below @xmath3 even in the most general mssm scenario , and at least two sparticles accessible to a @xmath108 linear collider . r\|\|c\|c\|c\| & general & gaugino & + dark + & mssm & unification & matter + @xmath24 & 140 _ ( ) _ & 135 _ ( ) _ & 105 _ ( ) _ + 5 & 165 _ ( 75 ) _ & 160 _ ( 65 ) _ & 125 _ ( 65 ) _ + 10 & 215 _ ( 135 ) _ & 210 _ ( 105 ) _ & 140 _ ( 105 ) _ + 30 & 335 _ ( 255 ) _ & 285 _ ( 205 ) _ & 260 _ ( 205 ) _ + 50 & & * 345 * _ ( 265 ) _ & * 345 * _ ( 265 ) _ + in the previous subsection , we derived bounds on the lightest sparticles , independent of the identity of those sparticles . another important piece of information that can be provided by this analysis is bounds on individual species of sparticles , for example , on the charginos or on the smuons . these bounds will of mathematical necessity be higher than those derived in the previous section , but still provide important information about how and where to look for susy . in particular , they can help us gauge the likelihood of finding susy at run ii of the tevatron or at the lhc . there is one complication in obtaining these bounds . at low @xmath2 the data is most easily explained by the neutralino diagrams and as such there must be at least one light smuon and one light neutralino . at larger @xmath2 values of @xmath47 as large as the e821 central value generally require contributions from the chargino diagrams , so there must be a light chargino and a light sneutrino . however the correlations already discussed preserve the bounds on the various species over the whole range of @xmath2 . a bound on @xmath109 implies a bound on @xmath110 , and a bound on @xmath111 implies a bound on at least one of the @xmath112 , and in certain cases ( such as gaugino unification ) , the converses may be true as well . we have shown in fig . [ species ] the mass bounds on @xmath113 and @xmath114 under the assumption of gaugino unification ; a plot for @xmath115/@xmath116 will appear later in our discussion of tevatron physics . note that @xmath114 must be relatively light , even for large @xmath2 , thanks to the gaugino unification condition , while @xmath113 can be heavier but must still lie below @xmath3 at @xmath1 . finally , we can consider the general mssm without gaugino unification . the results can best be summarized by showing the mass bounds on the various sparticles at @xmath100 in relations to their bounds in the unified case . this is done in fig . [ nogut ] . = 4.5 in we can see from the figure that the bound on the @xmath113 is essentially identical to that in the gaugino unification picture . however the gaugino masses have shifted , and the reason is no mystery . once again , as discussed in the previous section , the lightest neutralino is no longer a @xmath57-like spectator to the magnetic moment , but is a @xmath117-like partner of a participating @xmath117-like chargino . we summarize our results for the various sparticle species in table [ masseachtable ] . there we have shown the upper bounds on several sparticles in the general mssm , the mssm with gaugino unification , and the previous case with the additional requirement of a neutralino lsp ( `` dark matter '' ) . the bounds represents those obtained using the @xmath1 limit ( central value ) of the e821 data . c\|\|c\|c\|c\| & general & gaugino & + dark + & mssm & unification & matter + @xmath114 & 610 _ ( 455 ) _ & 470 _ ( 350 ) _ & 340 _ ( 265 ) _ + @xmath116 & none _ ( none ) _ & 830 _ ( 565 ) _ & 575 _ ( 440 ) _ + @xmath115 & none _ ( 690 ) _ & 830 _ ( 565 ) _ & 575 _ ( 440 ) _ + @xmath113 & 870_(680 ) _ & 825 _ ( 665 ) _ & 825 _ ( 665 ) _ + @xmath118 & 865 _ ( 675 ) _ & 820 _ ( 660 ) _ & 820 _ ( 660 ) _ + the final bound we will derive using the e821 data is on @xmath2 . there has been some discussion in the literature about which values of @xmath2 are capable of explaining the data . and in fact , we concur that at lower @xmath2 , there is a real suppression in the maximum size of @xmath47 . so in figure [ tanb ] we have shown the maximum attainable value of @xmath47 as a function of @xmath2 with and without the added assumption of gaugino mass unification . = 4.25 in the limit in fig . [ tanb ] clearly divides into two regions . at @xmath119 the chargino contribution dominates and thus @xmath120 , scaling linearly with @xmath2 . at lower @xmath47 , however , both neutralino and chargino contributions can be important so it becomes possible to generate @xmath47 with much smaller values of @xmath2 than would be possible from the charginos alone . the central value of the e821 data implies @xmath5 ; but already at @xmath1 all values of @xmath2 down to 1 are allowed . this result contradicts most statements made in the literature to date about limiting @xmath2 using @xmath0 : at present we are not able to place any @xmath1 bound on @xmath2 larger than 1 due to the very real possibility of a neutralino - dominated @xmath47 . future reduction in the error bars on @xmath0 will provide an important opportunity for placing a meaningful lower bound on @xmath2 . at its simplest level , the measurement of @xmath47 , an anomaly in the lepton sector , has little impact on the tevatron , a hadron machine . in particular , the light smuons associated with @xmath121 can not be directly produced at the tevatron , occuring only if heavier ( non - leptonic ) states are produced which then decay to sleptons . in the calculation of @xmath0 , the only such sparticles are the neutralinos and charginos . these states can be copiously produced and in fact form the initial state for the `` gold - plated '' susy trilepton signature . of particular interest for the trilepton signature are the masses of the lighter chargino ( @xmath115 ) and 2nd lightest neutralino ( @xmath116 ) . studies of msugra parameter space indicate that the sensitivity to the trilepton signature at run ii / iii of the tevatron depends strongly on the mass of sleptons which can appear in the gaugino decay chains . for heavy sleptons , the tevatron is only sensitive to gaugino masses in the range @xcite @xmath122 to @xmath123 for @xmath124 of luminosity and @xmath125 to @xmath126 for @xmath127 , with smaller values for smaller @xmath2 . however , for light sleptons ( below about @xmath128 ) the range is considerably extended , up to gaugino masses around 190 to @xmath129 . of course , this latter range is exactly the one most relevant for understanding @xmath0 . it is impossible in the kind of analysis presented here to comment on the expected cross - sections for the neutralino - chargino production ( there is no information in @xmath0 on the masses of the @xmath130-channel squarks , for example ) but we can examine the mass bounds on @xmath115 and @xmath116 . in fig . [ tevatron ] we have shown just that : the upper bound on the _ heavier _ of either @xmath115 or @xmath116 as a function of @xmath47 for several values of @xmath2 . a few comments are in order on the plot . first , the plot assumes gaugino unification ; dropping that assumption can lead to significantly heavier masses for the @xmath116 though not for the @xmath115 ; this is because the light @xmath115 is needed to participate in the magnetic moment , while @xmath116 can decouple . second , we have also assumed a neutralino lsp ; this is to be expected since the event topology for the trilepton signal assumes a stable , neutralino lsp . finally , on the @xmath131-axis is actually plotted @xmath111 , but in every case we examined with gaugino unification , the difference in the maximum masses of @xmath115 and @xmath116 differed by at most a few gev . this is because they are both dominantly wino - like in the unified case and thus have masses @xmath132 . = 4.25 in from the figure it is clear that a no - lose theorem for the tevatron is not lurking in the current e821 data . however , if the central value reported by e821 holds up and @xmath133 then one should expect the tevatron to find a trilepton signal for susy . so there is actually significant hope for a positive result . we can not emphasize enough too that these are _ upper bounds _ on the sparticle masses and in no way represent best fits or preferred values . thus even for larger @xmath2 or smaller @xmath47 , there is good reason to hope that the tevatron will be able to probe the gaugino sector in run ii or iii . a concensus is currently forming in favor of building a @xmath134 linear collider , presumably a factory for sparticles with masses below @xmath135 . what does the measurement of @xmath0 tell us with regards to our chances for seeing susy at @xmath134 ? and how many sparticles will be actually accessible to such a collider ? the analysis of the previous section can put a lower bound on the number of observable sparticles at a linear collider as a function of @xmath47 and @xmath2 and we show those numbers as a histogram in fig . [ nlchist ] . in this figure , we have shown the _ minimum _ number of sparticles with mass below @xmath135 for @xmath136 , 10 , 30 and 50 , assuming gaugino unification . in the graph , the thinner bars represent smaller @xmath2 . as is to be expected , the number of light states increases with increasing @xmath47 and with decreasing @xmath2 . however note that there are no @xmath136 lines for @xmath137 since there is no way to explain such large @xmath47 values at low @xmath2 . = 4.25 in we see from the figure also that for @xmath138 there is no guarantee that a @xmath59 machine would produce on - shell sparticles ; this is not to be taken to mean that one should not expect their production , simply that @xmath0 can not guarantee it . however for @xmath133 , the @xmath1 limit on @xmath0 from e821 indicates that at least 2 to 3 sparticles will be accessible to a @xmath59 machine . this counting does not include extra sleptons due to slepton mass universality ; for example , a light muon sneutrino also implies light tau and electron sneutrinos , and likewise for the charged smuon . a similar bar graph can also be made for a @xmath139 machine , though we do not show it here . however the relevant numbers can be inferred from fig . [ massfig ] ; we see that such a machine is guaranteed to produce at least three sparticles for @xmath103 at the e821 @xmath1 limit . deviations in the muon anomalous magnetic moment have long been advertised as a key hunting ground for indirect signatures of susy . and it may be that we finally have the evidence we need in the @xmath13 deviation reported by e821 . if this signal is real and if susy is the correct explanation for it , then light sparticles are requisite . in this analysis , we have confirmed , extended and overridden some of the bounds on these light sparticles that have appeared previously . in particular , for the @xmath1 lower bound on @xmath47 derived from the e821 data , we obtain the following for the most general mssm parametrization : * there must be at least 4 sparticles with masses below @xmath3 ( @xmath22 for unified gauginos ) ; * the lightest sparticle must lie below @xmath4 ( @xmath140 for unified gauginos ) ; * there is no lower bound on @xmath2 ( the central value of the e821 data requires @xmath5 ) ; * the @xmath113 and the @xmath118 must lie below @xmath141 while one neutralino must fall below @xmath142 ( those bounds become @xmath143 and @xmath144 respectively for unified gauginos ) ; * there is no upper bound on @xmath111 ( the data s central value requires @xmath6 ) ; * if @xmath133 , then a @xmath59 linear collider is guaranteed to produce at least 2 sparticles , and run ii / iii of the tevatron is likely to see a trilepton signature ( assuming a stable neutralino lsp ) ; for @xmath145 , neither of these can be guaranteed . in the most interesting theoretical scenario , gaugino unification with a neutralino lsp , the bounds on susy masses can be much lower , with the neutralino lsp below @xmath146 , @xmath115 and @xmath116 below @xmath147 and sleptons below @xmath148 . our only important simplifications were to drop the dependence on @xmath68 from the smuon mixing matrices , and to assume slepton mass universality . if e821 has discovered evidence for susy , then it is clear that we have many more discoveries awaiting us . the current data implies that the lhc will find susy directly ; such a discovery for the tevatron or the nlc is not guaranteed , though the data on @xmath0 greatly enhances the likelihood that these programs will be successful . we are grateful to s. martin for discussions during the early parts of this project . ck would also like to thank the aspen center for physics where parts of this work were completed , and the notre dame high - performance computing cluster for much - needed computing resources . this work was supported in part by the national science foundation under grant nsf-0098791 . the work of mb was supported in part by a notre dame center for applied mathematics graduate summer fellowship . the supersymmetric contributions to @xmath0 are generated by diagrams involving charged smuons with neutralinos , and sneutrinos with charginos . the most general form of the calculation , including phases , takes the form ( we follow ref . @xcite ) : a_^ ( ) & = & _ i , m\ { - ( \|n^l_im\|^2 + \|n^r_im\|^2)f^n_1(x_im ) + [ n^l_imn^r_im ] f^n_2(x_im ) } + a_^ ( ) & = & _ k \ { ( \|c^l_k\|^2 + \|c^r_k\|^2)f^c_1(x_k ) + [ c^l_kc^r_k ] f^c_2(x_k)}where @xmath149 , @xmath150 , and @xmath151 label the neutralino , smuon and chargino mass eigenstates respectively , and n^r_im & = & g_1 n_i1 x_m2 + y_n_i3 x_m1 , + n^l_im & = & 1 ( g_2 n_i2 + g_1 n_i1 ) x_m1^ * - y_n_i3 x^_m2 , + c^r_k & = & y_u_k2 , + c^l_k & = & -g_2v_k1 , @xmath152 is the muon yukawa coupling , and @xmath153 are the u(1 ) hypercharge and su(2 ) gauge couplings . the loop functions depend on the variables @xmath154 , @xmath155 and are given by f^n_1(x ) & = & + f^n_2(x ) & = & + f^c_1(x ) & = & + f^c_2(x ) & = & -for degenerate sparticles ( @xmath156 ) the functions are normalized so that @xmath157 . we can also bound the magnitude of some of these functions ; in particular @xmath158 while @xmath159 is unbounded as @xmath160 . ) [ charginomassmatrix ] where @xmath161 , @xmath162 and likewise for @xmath163 . the neutralino mixing matrix @xmath164 and the chargino mixing matrices @xmath165 and @xmath166 satisfy n^ m _ n^&= & diag ( m__1 , m__2 , m__3 , m__4 ) + u^ * m _ v^&= & diag ( m__1 , m__2 ) . ) where m_l^2 & = & m^2_l + ( s_w^2 -)m_z^22 + m_r^2 & = & m^2_r -s_w^2 m_z^22 for soft masses @xmath168 and @xmath169 ; the unitary smuon mixing matrix @xmath170 is defined by x m^2 _ x^= diag ( m^2__1 , m^2__2 ) . we will define a smuon mixing angle @xmath171 such that @xmath172 and @xmath173 . in our numerical calculations we will set @xmath174 . at low @xmath2 we have checked that varying @xmath68 makes only slight numerical difference , while at large @xmath2 it has no observable effect whatsoever . finally , the muon sneutrino mass is related to the left - handed smuon mass parameter by m_^2 = m_l^2 + 12 m_z^2 2 . the leading 2-loop contributions to @xmath47 have been calculated @xcite and have been found to suppress the susy contribution by a factor @xmath175 ; we will include this 7% suppression in all our numerical results . h. n. brown _ et al . _ [ muon g-2 collaboration ] , phys . lett . * 86 * , 2227 ( 2001 ) . a. czarnecki and w. j. marciano , phys . d * 64 * , 013014 ( 2001 ) . l. everett , g. kane , s. rigolin and l. wang , phys . lett . * 86 * , 3484 ( 2001 ) . j. l. feng and k. t. matchev , phys . * 86 * , 3480 ( 2001 ) . s. p. martin and j. d. wells , phys . d * 64 * , 035003 ( 2001 ) . u. chattopadhyay and p. nath , phys . * 86 * , 5854 ( 2001 ) ; + s. komine , t. moroi and m. yamaguchi , phys . lett . b * 506 * , 93 ( 2001 ) and phys . b * 507 * , 224 ( 2001 ) ; + j. r. ellis , d. v. nanopoulos and k. a. olive , phys . b * 508 * , 65 ( 2001 ) ; + r. arnowitt , b. dutta , b. hu and y. santoso , phys . b * 505 * , 177 ( 2001 ) ; + k. choi , phys . rev . d * 64 * , 055001 ( 2001 ) ; + j. e. kim , b. kyae and h. m. lee , hep - ph/0103054 ; + k. cheung , c. chou and o. c. kong , hep - ph/0103183 ; + h. baer , phys . rev . d * 64 * , 035004 ( 2001 ) ; + f. richard , hep - ph/0104106 ; + c. chen and c. q. geng , phys . b * 511 * , 77 ( 2001 ) ; + k. enqvist , e. gabrielli and k. huitu , phys . b * 512 * , 107 ( 2001 ) ; + d. g. cerdeno , hep - ph/0104242 ; + g. cho and k. hagiwara , phys . b * 514 * , 123 ( 2001 ) . m. davier and a. hcker , phys . b * 435 * , 427 ( 1998 ) . f. j. yndurain , hep - ph/0102312 . w. j. marciano and b. lee roberts , hep - ph/0105056 . j. a. grifols and a. mendez , phys . d * 26 * , 1809 ( 1982 ) ; + r. barbieri and l. maiani , phys . b * 117 * , 203 ( 1982 ) ; + d. a. kosower , l. m. krauss and n. sakai , phys . b * 133 * , 305 ( 1983 ) ; + t. c. yuan , z. phys . c * 26 * , 407 ( 1984 ) ; + u. chattopadhyay and p. nath , phys . rev . d * 53 * , 1648 ( 1996 ) ; + t. moroi , phys . d * 53 * , 6565 ( 1996 ) , [ * e * : _ ibid . _ d * 56 * , 4424 ( 1996 ) ] ; + m. carena , g. f. giudice and c. e. wagner , phys . b * 390 * , 234 ( 1997 ) ; + t. ibrahim and p. nath , phys . d * 61 * , 095008 ( 2000 ) . h. e. haber and g. l. kane , phys . rept . * 117 * , 75 ( 1985 ) . for a review , see : + s. abel _ et al . _ [ tevatron run ii sugra working group ] , hep - ph/0003154 . g. degrassi and g. f. giudice , phys . d * 58 * , 053007 ( 1998 ) .	we re - examine the bounds on supersymmetric particle masses in light of the e821 data on the muon anomalous magnetic moment , @xmath0 . we confirm , extend and supersede previous bounds . in particular we find ( at @xmath1 ) no lower limit on @xmath2 or upper limit on the chargino mass implied by the data at present , but at least 4 sparticles must be lighter than 700 to @xmath3 and at least one sparticle must be lighter than 345 to @xmath4 . however , the e821 central value bounds @xmath5 and @xmath6 . for @xmath7 , the data indicates a high probability for direct discovery of susy at run ii or iii of the tevatron . @=11 caption#1[#2]#3 @=12 .7ex .7ex gev mev ev kev tev m_z m_pl _ _ fc ^-1 ^-1 * d * * d^ * * u * * u^ * * y_d * * y_d^ * * y_u * * y_u^ * * v * * v^ * * v^0 * * v^0 * * x * b^0-\|b^0 # 1#1 \| # 1\| # 1 * m * m_susy m_unif # 1#2#3nucl . phys . * b#1 * ( 19#2 ) # 3 # 1#2#3phys . lett . * b#1 * ( 19#2 ) # 3 # 1#2#3phys . lett . * # 1b * ( 19#2 ) # 3 # 1#2#3phys . rev . * d#1 * ( 19#2 ) # 3 # 1#2#3phys . rev . lett . * # 1 * ( 19#2 ) # 3 # 1#2#3phys . rep . * # 1 * ( 19#2 ) # 3 # 1#2#3ann . rev . astron . astrophys . * # 1 * ( 19#2 ) # 3 # 1#2#3ann . rev . nucl . part . sci . * # 1 * ( 19#2 ) # 3 # 1#2#3mod . . lett . * # 1 * ( 19#2 ) # 3 # 1#2#3zeit . fr physik * c#1 * ( 19#2 ) # 3 # 1#2#3ap . j. * # 1 * ( 19#2 ) # 3 # 1#2#3ann . phys . * # 1 * ( 19#2 ) # 3 # 1#2#3rev . mod . phys . * # 1 * ( 19#2 ) # 3 # 1#2#3comm . math . phys . * # 1 * ( 19#2 ) # 3 _ i.e. _ _ et al . _ _ e.g. _ _ etc . _ _ c.f . _ /#1/ \|#1 # 1#1 und - hep-01-k03 + august 2001 + * implications of the muon anomalous magnetic moment + for supersymmetry * + 0.5 cm _ department of physics , university of notre dame + notre dame , in 46556 , usa + _ .5 cm recently , the brookhaven e821 experiment @xcite has reported evidence for a deviation of the muon magnetic moment from the standard model expectation . immediately following that announcement appeared a number of papers analyzing the reported excess in terms of various forms of new physics , including supersymmetry ( susy ) @xcite . the susy explanation is , for many of us , the most exciting of the various proposals since it implies susy at a mass scale not far above the weak scale . in particular , it implies a light slepton and a light gaugino , though `` light '' can still be as heavy as several hundred gev . previous analyses have generally concentrated on bounding the masses of the chargino and the sneutrino , or on the lightest of the sparticles . in this paper , we will re - examine the susy calculation , both in very general scenarios and in well - motivated , simplified susy models . we will show that there exist mass bounds on more than two susy particles , though which sparticles are bounded changes with @xmath2 . in particular , we will show that throughout parameter space , one may actually bound _ at least 4 _ sparticle masses if we are to use susy to explain the e821 data . the existence of these bounds will rely on very simple and clearly stated assumptions about the susy particle spectrum ; these assumptions will not include a fine - tuning constraint .
[ sec:1 ] magnetized plasmas are encountered in a wide variety of astrophysical situations , but also in magnetic fusion devices such as tokamaks , where a large external magnetic field needs to be applied in order to keep the particles on the desired tracks . in particle - in - cell ( pic ) simulations of such devices , this large external magnetic field obviously needs to be taken into account when pushing the particles . however , due to the magnitude of the concerned field this often adds a new time scale to the simulation and thus a stringent restriction on the time step . in order to get rid of this additional time scale , we would like to find approximate equations , where only the gross behavior implied by the external field would be retained and which could be used in a numerical simulation . in the simplest situation , the trajectory of a particle in a constant magnetic field @xmath0 is a helicoid along the magnetic field lines with a radius proportional to the inverse of the magnitude of @xmath0 . hence , when this field becomes very large the particle gets trapped along the magnetic field lines . however due to the fast oscillations around the apparent trajectory , its apparent velocity is smaller than the actual one . this result has been known for some time as the guiding center approximation , and the link between the real and the apparent velocity is well known in terms of @xmath0 . here , we consider a plasma constituted of a large number of charged particles , which is described by the vlasov equation coupled with the maxwell or poisson equations to compute the self - consistent fields . it describes the evolution of a system of particles under the effects of external and self - consistent fields . the unknown @xmath1 , depending on the time @xmath2 , the position @xmath3 , and the velocity @xmath4 , represents the distribution of particles in phase space for each species with @xmath5 , @xmath6 . its behaviour is given by the vlasov equation , @xmath7 where the force field @xmath8 is coupled with the distribution function @xmath9 giving a nonlinear system . we first define @xmath10 the charge density and @xmath11 the current density which are given by @xmath12 where @xmath13 is the elementary charge . for the vlasov - poisson model @xmath14 where @xmath15 represents the mass of one particle . on the other hand for the vlasov - maxwell model , we have @xmath16 and @xmath17 , @xmath0 are solutions of the maxwell equations @xmath18 with the compatibility condition @xmath19 which is verified by the solutions of the vlasov equation . here we will consider an intermediate model where the magnetic field is given , @xmath20 with @xmath21 and we focus on the long time behavior of the plasma in the orthogonal plane to the external magnetic field , that is the two dimensional vlasov - poisson system with an external strong magnetic field @xmath22 here , for simplicity we set all physical constants to one and consider that @xmath23 is a small parameter related to the ratio between the reciprocal larmor frequency and the advection time scale . the term @xmath24 in front of the time derivative of @xmath9 stands for the fact that we want to approximate the solution for large time . we want to construct numerical solutions to the vlasov - poisson system ( [ eq : vlasov2d ] ) by particle methods ( see @xcite ) , which consist in approximating the distribution function by a finite number of macro - particles . the trajectories of these particles are computed from the characteristic curves corresponding to the vlasov equation @xmath25 where the electric field is computed from a discretization of the poisson equation in ( [ eq : vlasov2d ] ) on a mesh of the physical space . the main purpose of this work is the construction of efficient numerical methods for stiff transport equations of type ( [ eq : vlasov2d ] ) in the limit @xmath26 . indeed , setting @xmath27 the system ( [ traj:00 ] ) can be re - written for @xmath28 as @xmath29 therefore , we denote by @xmath30 the solution to ( [ traj:01 ] ) , and under some classical smoothness assumptions on the electromagnetic fields @xmath31 , it is well - known at least when @xmath32 or @xmath33 is homogeneous that @xmath34 converges weakly to zero when @xmath35 , and @xmath36 , where @xmath37 corresponds to the guiding center approximation @xmath38 here , we are of course interested in the behavior of the sequence solution @xmath39 to the vlasov - poisson system ( [ eq : vlasov2d ] ) when @xmath26 , which corresponds to the gyro - kinetic approximation of the vlasov - poissons system . following the work of l. saint - raymond @xcite , it can be proved at least when @xmath33 is homogeneous that the charge density @xmath40 converges to the solution to the guiding center approximation @xmath41 -\delta _ { { \mathbf{x } } _ \perp}\phi=\rho , \end{array } \right . \label{eq : gc}\ ] ] where the velocity @xmath42 is @xmath43 we observe that the limit system ( [ traj : limit ] ) corresponds to the characteristic curves to the limit equation ( [ eq : gc ] ) . we seek a method that is able to capture these properties , while the numerical parameters may be kept independent of the stiffness degree of these scales . this concept is known and widely studied for dissipative systems in the framework of asymptotic preserving schemes @xcite . contrary to collisional kinetic equations in hydrodynamic or diffusion asymptotic , collisionless equations like the vlasov - poisson system ( [ eq : vlasov2d ] ) involve time oscillations . in this context , the situation is more complicated than the one encountered in collisional regimes since we can not expect any dissipative phenomenon . therefore , the notion of two - scale convergence has been introduced both at the theoretical and numerical level @xcite in order to derive asymptotic models . however , these asymptotic models , obtained after removing the fast scales , are valid only when @xmath24 is small . we refer to e. frnod and e. sonnendrcker @xcite , and f. golse and l. saint - raymond @xcite for a theoretical point of view on these questions , and e. frnod , f. salvarani and e. sonnendrcker @xcite for numerical applications of such techniques . another approach is to combine both disparate scales into one and single model . such a decomposition can be done using a micro - macro approach ( see @xcite and the references therein ) . such a model may be used when the small parameter of the equation is not everywhere small . hence , a scheme for a micro - macro model can switch from one regime to another without any treatment of the transition between the different regimes . a different method consists in separating fast and slow time scales when such a structure can be identified @xcite or @xcite . theses techniques work well when the magnetic field is uniform since fast scales can be computed from a formal asymptotic analysis , but for more complicated problems , that is , when the external magnetic field depends on time and position @xmath3 , the generalization of this approach is an open problem . in this paper , we propose an alternative to such methods allowing to make direct simulations of systems ( [ eq : vlasov2d ] ) with large time steps with respect to @xmath24 . we develop numerical schemes that are able to deal with a wide range of values for @xmath24 , so - called asymptotic preserving ( ap ) class @xcite , such schemes are consistent with the kinetic model for all positive value of @xmath24 , and degenerate into consistent schemes with the asymptotic model when @xmath44 . before presenting our time discretization technique , let us first briefly review the basic tools of particle - in - cell methods which are widely used for plasma physics simulations @xcite . the numerical resolution of the vlasov equation and related models is usually performed by particle - in - cell ( pic ) methods which approximate the plasma by a finite number of particles . trajectories of these particles are computed from characteristic curves ( [ traj:00 ] ) corresponding to the the vlasov equation ( [ eq : vlasov2d ] ) , whereas self - consistent fields are computed on a mesh of the physical space . this method yields satisfying results with a relatively small number of particles but it is sometimes subject to fluctuations , due to the numerical noise , which are difficult to control . to improve the accuracy , direct numerical simulation techniques have been developed . the vlasov equation is discretized in phase space using either semi - lagrangian @xcite , finite difference @xcite or discontinuous galerkin @xcite schemes . but these direct methods are very costly , hence several variants of particle methods have been developed over the past decades . in the complex particle kinetic scheme introduced by bateson and hewett @xcite , particles have a gaussian shape that is transformed by the local shearing of the flow . moreover they can be fragmented to probe for emerging features , and merged where fine particles are no longer needed . in the cloud in mesh ( cm ) scheme of alard and colombi @xcite particles also have gaussian shapes , and they are deformed by local linearization of the force field . more recently in @xcite , the authors proposed a linearly - transformed particle - in - cell method , that employs linear deformations of the particles . here we focus on the time discretization technique , hence we will only consider standard particle method even if our approach is completely independant on the choice of the particle method . the particles method consists in approximating the initial condition @xmath45 in ( [ eq : vlasov2d ] ) by the following dirac mass sum @xmath46 where @xmath47 is a beam of @xmath48 particles distributed in the four dimensional phase space according to the density function @xmath45 . afterwards , one approximates the solution of ( [ eq : vlasov2d ] ) , by @xmath49 where @xmath50 is the position in phase space of particle @xmath51 moving along the characteristic curves ( [ traj:00 ] ) with the initial data @xmath52 , for @xmath53 . however when the vlasov equation is coupled with the poisson equation for the computation of the electric field , the dirac mass has to be replaced by a smooth function @xmath54 @xmath55 where @xmath56 is a particle shape function with radius proportional to @xmath24 , usually seen as a smooth approximation of the dirac measure obtained by scaling a compactly supported `` cut - off '' function @xmath57 for which common choices include b - splines and smoothing kernels with vanishing moments , see _ e.g. _ @xcite . particle centers are then pushed forward at each time @xmath2 by following a numerical approximation of the flow ( [ traj:00 ] ) , leading to @xmath58 in the classical error analysis @xcite , the above process is seen as * an approximation ( in the distribution sense ) of the initial data by a collection of weighted dirac measures ; * the exact transport of the dirac particles along the flow ; * the smoothing of the resulting distribution @xmath59 with the convolution kernel @xmath60 . the classical error estimate reads then as follows @xcite : [ lmm:0 ] consider the vlasov equation with a given electromagnetic field @xmath61 and a smooth initial datum @xmath62 , with @xmath63 . if for some prescribed integers @xmath64 and @xmath65 , the cut - off @xmath66 has @xmath15-th order smoothness and satisfies a moment condition of order @xmath67 , namely , @xmath68 and @xmath69 then there exists a constant @xmath70 independent of @xmath45 , @xmath48 or @xmath71 , such that we have for all @xmath72 , @xmath73 when @xmath74 and @xmath75 where the ratio @xmath76 . note that following @xcite , it is also possible to get explicit order of convergence for the linear vlasov equation . let us also mention related papers where the convergence of a numerical scheme for the vlasov - poisson system is investigated . cottet and p .- a . raviart @xcite present a precise mathematical analysis of the particle method for solving the one - dimensional vlasov poisson system . we also mention the papers of s. wollman and e. ozizmir @xcite and s. wollman @xcite on the topic . k. ganguly and h.d . victory give a convergence result for the vlasov - maxwell system @xcite . the rest of the paper is organized as follows . in section [ sec:3 ] we present several time discretization techniques based on high - order semi - implicit schemes @xcite for the vlasov - poisson system with a strong external magnetic field , and we prove uniform consistency of the schemes in the limit @xmath35 with preservation of the order of accuracy ( from first to third order accuracy ) . in section [ sec:4 ] we perform a rigorous analysis of the first order scheme for smooth electromagnetic fields . section [ sec:5 ] is then devoted to numerical simulations for one single particle motion and for the vlasov - poisson model for various asymptotics @xmath77 and @xmath78 , which illustrate the advantage of high order schemes . [ sec:3 ] let us now consider the system ( [ eq : vlasov2d ] ) and apply a particle method , where the key issue is to design a uniformly stable scheme with respect to the parameter @xmath79 , which is related to the magnitude of the external magnetic field . assume that at time @xmath80 , the set of particles are located in @xmath81 , we want to solve the following system on the time interval @xmath82 $ ] , @xmath83 where the electric field is computed from a discretization of the poisson equation ( [ eq : vlasov2d ] ) on a mesh of the physical space . the numerical scheme that we describe here is proposed in the framework of particle - in - cell method , where the solution @xmath9 is discretized as follows @xmath84 where @xmath85 represents an approximation of the solution @xmath86 and @xmath87 to ( [ traj : bis ] ) . when the vlasov equation ( [ eq : vlasov2d ] ) is coupled with the poisson equation ( [ poisson ] ) , the electric field is computed in a macro - particle position @xmath88 at time @xmath89 as follows * compute the density @xmath90 @xmath91 * solve a discrete approximation to ( [ poisson ] ) @xmath92 * interpolate the electric field with the same order of accuracy on the points @xmath93 . to discretize the system ( [ traj : bis ] ) , we apply the strategy developed in @xcite based on semi - implicit solver for stiff problems . in the rest of this section , we propose several numerical schemes to the system ( [ traj : bis ] ) for which the index @xmath94 will be omitted . we start with the simplest semi - implicit scheme for ( [ traj : bis ] ) , which is a combination of backward and forward euler scheme . it gives for a fixed time step @xmath95 and a given electric field @xmath96 and an external magnetic field @xmath33 , @xmath97 note that only the second equation on @xmath98 is fully implicit and requires the inversion of a linear operator . then , from @xmath98 the first equation gives the value for the position @xmath99 . [ prop:1 ] let us consider a time step @xmath95 , a final time @xmath100 and set @xmath101 $ ] . assume that the sequence @xmath102 given by ( [ scheme:0 ] ) is such that for all @xmath103 , @xmath104 is uniformly bounded with respect to @xmath79 and @xmath105 converges in the limit @xmath35 to some @xmath106 . then , for @xmath103 , @xmath107 , as @xmath35 and the limit @xmath108 is a consistent first order approximation with respect to @xmath109 of the guiding center equation provided by the scheme @xmath110 for all @xmath103 , we consider @xmath111 the solution to ( [ scheme:0 ] ) now labeled with respect to @xmath79 . since , the sequence @xmath112 is uniformly bounded with respect to @xmath79 , we can extract a subsequence still abusively labeled by @xmath113 and find some @xmath114 such that @xmath107 as @xmath113 goes to zero . then , we observe that the second equation of ( [ scheme:0 ] ) can be written as @xmath115 and that , for any @xmath116 , @xmath117 is uniformly bounded . from this we conclude first that for any @xmath103 , @xmath118 is uniformly bounded then that @xmath119 for any @xmath116 . therefore , taking the limit @xmath35 , it yields that for @xmath120 , @xmath121 substituting the limit of @xmath122 in the first equation of ( [ scheme:0 ] ) we prove that the limit @xmath123 satisfies ( [ sch : y0 ] ) . since the limit point @xmath124 is uniquely determined , actually all the sequence @xmath112 converges . the consistency provided by the latter result is far from being uniform with respect to the time step . however we do prove in the next section that the solution to ( [ scheme:0 ] ) is both uniformly stable and consistent with respect to @xmath109 and @xmath79 . of course , such a first order scheme is not accurate enough to describe correctly the long time behavior of the numerical solution , but it has the advantage of the simplicity and we will prove in the next section that it is uniformly stable with respect to the parameter @xmath113 and the sequence @xmath125 converges to a consistent approximation of the guiding center model when @xmath35 . now , let us see how to generalize such an approach to second and third order schemes . now , we consider second order schemes with two stages . a first example of scheme satisfying the second order conditions is given by a combination of heun method ( explicit part ) and an @xmath126-stable second order singly diagonal implicit runge - kutta sdirk method ( implicit part ) @xcite . the first stage corresponds to @xmath127 . } \end{array}\right.\ ] ] then the second stage is given by @xmath128 . } \end{array}\right.\ ] ] finally , the numerical solution at the new time step is @xmath129 a similar numerical scheme has been proposed in the framework of @xmath130 simulation of the vlasov - poisson system @xcite . under stability assumptions on the numerical solution to ( [ scheme:2 - 1])-([scheme:2 - 3 ] ) , we get the following consistency result in the limit @xmath35 . [ prop:2 ] let us consider a time step @xmath95 , a final time @xmath100 and set @xmath101 $ ] . assume that the sequence @xmath102 given by ( [ scheme:2 - 1])-([scheme:2 - 3 ] ) is such that for all @xmath103 , @xmath104 is uniformly bounded with respect to @xmath79 and @xmath105 converges in the limit @xmath35 to some @xmath106 . then , for @xmath103 , @xmath107 , as @xmath35 and the limit @xmath131 is a consistent and second order approximation with respect to @xmath109 of the guiding center equation given by the scheme @xmath132 and @xmath133 . we follow the lines of the proof of proposition [ prop:1 ] and first choose a subsequence of @xmath125 converging to some @xmath123 . the second equation of implies that @xmath134 is bounded . from the first equation of it follows that so is @xmath135 . then the second equation of yields boundedness of @xmath136 and the first that @xmath137 is also bounded . now from the second equation of stems that @xmath138 converges to zero as @xmath139 for any @xmath116 . coming back to the second equation of we show that for all @xmath140 , @xmath134 converges when @xmath35 and @xmath141 using the first equation of we conclude that @xmath142 also converges and that its limit @xmath143 is given by the first equation of . going on with the same arguments shows that @xmath144 and that @xmath145 converges to a @xmath146 given by the second equation of . one may then take a limit in the first equation of and conclude that indeed the limit @xmath123 satisfies ( [ sch : y1 ] ) . again uniqueness supplies the convergence of the whole sequence . another choice is a combination of runge - kutta method ( explicit part ) and an @xmath147-stable second order sdirk method in the implicit part . this implicit scheme should give better stability properties on the numerical solution with respect to the stiffness parameter @xmath79 . we first choose @xmath148 as the smallest root of the polynomial @xmath149 , _ i.e. _ @xmath150 , then the scheme is given by the following two stages . first , we have @xmath151 with @xmath152 for the second stage , we first define @xmath153 then the solution @xmath154 is given by @xmath155 with @xmath156 finally , the numerical solution at the new time step is @xmath157 under stability assumptions on the numerical solution to ( [ scheme:2 - 1])-([scheme:2 - 3 ] ) , we get the following consistency result in the limit @xmath35 . [ prop:3 ] let us consider a time step @xmath95 , a final time @xmath100 and set @xmath101 $ ] . assume that the sequence @xmath102 given by ( [ scheme:3 - 1])-([scheme:3 - 3 ] ) is such that for all @xmath103 , @xmath104 is uniformly bounded with respect to @xmath79 and @xmath105 converges in the limit @xmath35 to some @xmath106 . then , for @xmath103 , @xmath107 , as @xmath35 and the limit @xmath131 is a consistent second order approximation of the guiding center equation , given by @xmath158 where @xmath159 we omit the proof of proposition [ prop:3 ] as almost identical to the one of proposition [ prop:2 ] . the present scheme is @xmath147- stable , which means uniformly linearly stable with respect to @xmath109 . a third order semi - implicit scheme is given by a four stages runge - kutta method introduced in the framework of hyperbolic systems with stiff source terms @xcite . first , we set @xmath160 , @xmath161 and @xmath162 and @xmath163 . then we construct the first stage as @xmath164 with @xmath165 for the second stage , we have @xmath166 with @xmath167 then , for the third stage we set @xmath168 with @xmath169 finally , for the fourth stage we set @xmath170 with @xmath171 and the numerical solution at the new time step is @xmath172 as for the previous schemes , under uniform stability assumptions with respect to @xmath79 , we prove the following proposition [ prop:5 ] let us consider a time step @xmath95 , a final time @xmath100 and set @xmath101 $ ] . assume that the sequence @xmath102 given by ( [ scheme:4 - 1])-([scheme:4 - 5 ] ) is such that for all @xmath103 , @xmath104 is uniformly bounded with respect to @xmath79 and @xmath105 converges in the limit @xmath35 to some @xmath106 . then , for @xmath103 , @xmath107 , as @xmath35 and the limit @xmath131 is a consistent third order approximation of the guiding center equation provided by the scheme @xmath173 where @xmath174 and @xmath175 we follow the lines of the proof of proposition [ prop:2 ] . in particular we prove successively the boundedness of @xmath134 , @xmath135 , @xmath136 , @xmath137 , @xmath176 , @xmath177 , @xmath178 , @xmath179 . arguing inductively on @xmath140 one then shows that @xmath180 , @xmath181 , @xmath182 , @xmath183 , @xmath184 all converge to zero when @xmath139 . then for any converging subsequence of @xmath185 we may successively identify limits for @xmath134 , @xmath135 , @xmath136 , @xmath137 , @xmath176 , @xmath177 , @xmath178 , @xmath179 and prove that any accumulating point of @xmath185 solves the limiting scheme involving . hence the result . [ sec:4 ] consider the first order euler semi - implicit scheme @xmath186 we focus here on the case where both @xmath187 and @xmath33 are given external fields . since , in this case , there is no coupling between particles we may focus safely on one of them and drop the @xmath188 suffix . then we prove we set @xmath189 assume that the @xmath190 and consider the solution @xmath191 to ( [ traj : ter ] ) . then there exist positive constants @xmath70 and @xmath192 , only depending on @xmath187 , @xmath33 and @xmath193 such that when by any upper bound on @xmath109 but @xmath192 and @xmath70 would depend on this upper bound . ] @xmath194 and @xmath195 @xmath196\right\\| \right],\ ] ] where @xmath114 corresponds to the numerical solution of the guiding center model @xmath197 [ th:1 ] to start with we analyze the case where @xmath187 is a given bounded lipschitz field and @xmath33 is constant and of norm @xmath198 . our estimates shall be expressed in terms of @xmath199 for concision s sake we also introduce the operator to denote @xmath200=\ { { \mathbf w}}\wedge { \bf b}_{\rm ext}$ ] . by introducing the key quantity for @xmath201 , @xmath202,\ ] ] the first equation of reads @xmath203+\delta t\ { \mathbf{z } } ^n\,,\quad n\geq1\ ] ] while @xmath204^{-1}(\eps^{-1 } { \mathbf{v } } ^0-r[{{\mathbf e}}(t^0 , { \mathbf{x } } ^0)])$ ] and @xmath205^{-1 } ( { \mathbf{z } } ^n - r[{{\mathbf e}}(t^n , { \mathbf{x } } ^n)-{{\mathbf e}}(t^{n-1 } , { \mathbf{x } } ^{n-1})])\,,\quad n\geq1\,.\ ] ] we first observe that @xmath206 and @xmath207^{-1}=({{\rm id}}+\lambda r)/(1+\lambda^2)$ ] since @xmath208 . this leads to @xmath209\,,\quad n\geq1\,.\ ] ] assuming @xmath210 and introducing @xmath211 we infer @xmath212 where @xmath213\right\\|\,.\ ] ] for comparison we define @xmath123 solution to ( [ cg:01 ] ) . then , @xmath214,\quad n\geq1\,.\end{aligned}\ ] ] hence assuming moreover @xmath215 ( or replacing @xmath216 with some arbitrary positive number if @xmath217 ) @xmath218 \\ & + & \frac{b\,\delta t}{(1+k_x\delta t)[1- \frac{1+\lambda}{1+\lambda^2}]}\,\left[1+\left(\frac{1+\lambda}{1+\lambda^2}\right)^n\right]\,(1+k_x\delta t)^n\,.\end{aligned}\ ] ] as a conclusion , for any @xmath219 , there exists @xmath220 such that if @xmath221 then , it yields @xmath222\right\\|\right]\ e^{k_x\,n\delta t}\,,\ ] ] which concludes the proof of theorem [ th:1 ] when the magnetic field is uniform . we now relax the assumption that @xmath33 is constant and of norm @xmath198 . we set @xmath223 and let @xmath224 be dependent on @xmath2 and @xmath3 . we observe that now @xmath225 so that @xmath207^{-1}=({{\rm id}}+\lambda r)/(1+\lambda^2\mathbf{b}_{\rm ext}^2)$ ] . then introducing the drift force @xmath226 we essentially obtain the same estimates with @xmath227 replaced by @xmath228 and @xmath229 indeed introducing for @xmath201 , @xmath230 the scheme is written @xmath231^{-1}\left(\frac { { \mathbf{v } } ^0}{\eps}-{{\mathbf f}}(t^{0 } , { \mathbf{x } } ^{0})\right),\ ] ] and then @xmath232^{-1 } ( { \mathbf{z } } ^n-({{\mathbf f}}(t^{n } , { \mathbf{x } } ^{n})-{{\mathbf f}}(t^{n-1 } , { \mathbf{x } } ^{n-1})))\,,\quad n\geq1\ ] ] together with @xmath233 mark that if @xmath234 then @xmath235 note that this result can be slightly improved when we modify the initial condition of the asymptotic discrete model . indeed , consider @xmath123 solving @xmath236 the gain is that now @xmath237\ \left\\|\frac1\eps { \mathbf{v } } ^0-r[{{\mathbf e}}(t^0 , { \mathbf{x } } ^0)]\right\\|\ ] ] since @xmath238\right]\ -\ \frac1\lambda r\,[{{\rm id}}-\lambda r]^{-1}\left[\frac1\eps { \mathbf{v } } ^0-r[{{\mathbf e}}(t^0 , { \mathbf{x } } ^0)]\right]\,.\ ] ] this leads , for @xmath201 to @xmath239 \\ & + & \frac{b\,\delta t}{1- \frac{1+\lambda}{1+\lambda^2}}\,\frac{1+\lambda}{1+\lambda^2}\,\left[1+\left(\frac{1+\lambda}{1+\lambda^2}\right)^{n-1}\right]\,(1+k_x\delta t)^{n-1 } \\ & + & { \displaystyle } \frac{\delta t}{\lambda}\ \left[\,k_x\delta t\,+\,\frac{1+\lambda}{1+\lambda^2}\,\right]\ \left\\|\frac1\eps { \mathbf{v } } ^0-r[{{\mathbf e}}(t^0 , { \mathbf{x } } ^0)]\right\\|\ ( 1+k_x\delta t)^{n-1}\end{aligned}\ ] ] then with notation as above there exists a constant @xmath240 such that @xmath241\right\\|\right]\,.\ ] ] concerning the analysis of high - order schemes , it is not straightforward to adapt directly the strategy of theorem [ th:1 ] . indeed , the use of a semi - implicit scheme does not necessarily guarantee that the particle trajectories are under control . consider the scheme ( [ scheme:2 - 1])-([scheme:2 - 3 ] ) in the simplest situation where the electric field is zero and the magnetic field is @xmath242 , we show that the scheme preserves the kinetic energy and we have @xmath243 hence as @xmath113 goes to zero , the velocity @xmath244 can not converge to the null guiding center velocity except if the initial velocity does . fortunately , for the other high order schemes ( [ scheme:3 - 1])-([scheme:3 - 3 ] ) and ( [ scheme:4 - 1])-([scheme:4 - 5 ] ) , kinetic energy is dissipated and converges to 0 . [ rem : nc ] in this section , we discuss some examples to validate and to compare the different time discretization schemes . we first consider the single motion of a particle under the effect of a given electromagnetic field . it allows us to illustrate the ability of the semi - implicit schemes to capture the guiding center velocity with large time step @xmath245 in the limit @xmath26 . then we consider the vlasov - poisson system with an external magnetic field . a classical particle - in - cell method is applied with different time discretization techniques to compute the particle trajectories . hence this collection of charged particles move and give rise to a self - consistent electric field , obtained by solving numerically the poisson equation in a space grid . before going to the statistical descriptions , let us investigate the accuracy and stability properties of the semi - implicit algorithms presented in section [ sec:3 ] on the motion of individual particles in a given electromagnetic field . here we consider an electric field @xmath246 , where @xmath247 with @xmath248 and a magnetic field @xmath249 with @xmath250 . we choose for all simulations @xmath251 and the initial data as @xmath252 and @xmath253 , such that the initial data @xmath254 is bounded with respect to @xmath79 . [ cols="^,^ " , ] in this paper we proposed a class of semi - implicit time discretization techniques for particle - in cell simulations . the main feature of this approach is to guarantee the accuracy and stability when the amplitude of the magnetic field becomes large and to get the correct long time behavior ( guiding center approximation ) . we formally showed that the present schemes preserve the initial order of accuracy when @xmath35 . furthermore , we performed a complete analysis of the first order semi - implicit scheme when we consider a given and smooth electromagnetic field @xmath61 . the time discretization techniques proposed in this paper seem to be a very simple and efficient tool to filter fast oscillations and have nice stability and consistency properties in the limit @xmath35 . however , a complete analysis of high order semi - implicit schemes is still missing . the main issue is to control the space trajectory @xmath255 uniformly with respect to @xmath113 and as we have shown in remark [ rem : nc ] , the use of a semi - implicit scheme does not necessarily guarantee that the particle trajectories are under control . a complete analysis of high order schemes is currently under study . on the other hand , the present techniques will be applied to more advanced problems as the three dimensional vlasov - poisson system when the magnetic field is non uniform and the particle trajectories become more complicated . ff was supported by the eurofusion consortium and has received funding from the euratom research and training programme 2014 - 2018 under grant agreement no 633053 . the views and opinions expressed herein do not necessarily reflect those of the european commission . positivity preserving semi - lagrangian discontinuous galerkin formulation : theoretical analysis and application to the vlasov - poisson system , journal of computational physics , * 230 * , pp . 83868409 ( 2011 ) .	this paper deals with the numerical resolution of the vlasov - poisson system with a strong external magnetic field by particle - in - cell ( pic ) methods . in this regime , classical pic methods are subject to stability constraints on the time and space steps related to the small larmor radius and plasma frequency . here , we propose an asymptotic - preserving pic scheme which is not subjected to these limitations . our approach is based on first and higher order semi - implicit numerical schemes already validated on dissipative systems @xcite . additionally , when the magnitude of the external magnetic field becomes large , this method provides a consistent pic discretization of the guiding - center equation , that is , incompressible euler equation in vorticity form . we propose several numerical experiments which provide a solid validation of the method and its underlying concepts . keywords . high order time discretization ; vlasov - poisson system ; guiding - centre model ; particle methods .
in the present paper we consider a model where in a prescribed region @xmath0 of the euclidean space @xmath8 an agent ( a central government or a commercial company ) has the possibility to decide the price of a certain product ; this price @xmath1 may vary at each point @xmath2 and the customers density @xmath9 is assumed to be completely known . we assume that all the customers buy the same quantity of the product ; on the counterpart , a customer living at the point @xmath2 knows the pricing function @xmath3 everywhere and may decide to buy the product where he lives , then paying a cost @xmath1 , or in another place @xmath4 , then paying the cost @xmath5 and additionally a transportation cost @xmath6 for a given transportation cost function @xmath7 . the individual strategy of each customer is then to solve the minimization problem @xmath10 of particular importance to our problem is the ( set - valued ) map @xmath11 which associates to every customer living at the point @xmath12 all the locations where it is optimal to purchase the good . given the price pattern @xmath3 , @xmath13 is then defined by @xmath14 without any other constraint , due to the fact that the customers have to buy the product ( for instance gasoline , food , a medical product or cigarettes ) , the pricing strategy for the agent in order to maximize the total income would simply be increasing everywhere the function @xmath3 more and more . to avoid this trivial strategy we assume that on the region @xmath0 some kind of regulations are present , and we study the optimization problems the agent has to solve in order to maximize its total profit . we will study two models according to two different price constraints . we also assume that the supply is unconstrained at any location of the region @xmath0 , which means that whatever the total demand for the product is at a given location , it can be supplied by the agent to the customers . the simplest situation we consider is when the price @xmath1 is constrained to remain below a fixed bound @xmath15 everywhere on @xmath0 , due for instance to some regulatory policy . the only assumption we make is that @xmath16 is a _ proper _ nonnegative function , intending that in the region where @xmath17 no restrictions on @xmath1 are imposed . the goal of the agent is to maximize its total income that , with the notation introduced in and , can be written as @xmath18 under the constraint ( state equation ) that @xmath19 i.e. that @xmath20 is compatible with the customer s individual minimization problem . one may therefore see the previous program as a nonstandard optimal control problem where @xmath3 is the control and @xmath20 the state variable . let us mention that problems with a similar structure naturally arise in the so - called _ principal - agent _ problem in economics ( see for instance rochet and chon @xcite and the references therein ) . we consider a second model of pricing strategy : we suppose that in @xmath0 there is a given subregion @xmath21 where the price @xmath1 is fixed as a function @xmath15 that the agent can not control . this is for instance the case of another country if the agent represents a central government , or of a region where for some social reasons that the agent can not modify , the prices of the product are fixed . whenever @xmath22 , then the agent makes no benefit from customers living at @xmath12 . in fact the total profit of the agent is given by @xmath23 under the constraint ( state equation ) that @xmath19 . note that in formula giving the total profit , the integration is now performed only on the set of customers that do shop in the region controlled by the agent and not in the fixed - price region @xmath21 . the problem we are interested in reads again as the maximization of the functional @xmath24 among the admissible choices of state and control variables . for both models above we discuss the mathematical framework which enables us to obtain an existence result for an optimal pricing strategy and we present some particular cases where more detailed computations can be made , as the case of concave costs , the case of a quadratic cost , and the onedimensional case . the last section contains some discussions about possible extensions and developments , as for instance the case of nash equilibria when more agents operate on the same market . in what follows , @xmath0 will be some compact metric space ( the economic region ) , and @xmath25 $ ] a nonnegative proper function , i.e. we assume that @xmath16 is not @xmath26 everywhere on @xmath0 . we are also given a transportation cost function @xmath7 assumed to be continuous and nonnegative on @xmath27 and such that @xmath28 for all @xmath2 . finally , @xmath29 is a nonnegative radon measure on @xmath0 that models the distribution of customers in @xmath0 . the unknown of the problem is the pricing pattern @xmath3 that varies in the class @xmath30 once a price @xmath3 has been chosen by the agent , consumers living at any point @xmath2 purchase the good so as to minimize their total expenditure which is given by price plus commuting cost . this leads to the following definitions : @xmath31 by our l.s.c . and compactness assumptions , @xmath32 is a nonempty compact subset of @xmath0 ( but @xmath13 is not single - valued in general ) ; moreover the graph of @xmath13 is compact as the @xmath33 of some l.s.c . function on @xmath27 . note that @xmath32 is the set of locations where consumers living at @xmath12 rationally choose to purchase the good . it is possible however that , for some customers , the optimal total cost @xmath34 is reached at more than one point @xmath35 ; in this case we assume the _ tie - breaking rule _ that the consumers living at @xmath12 choose to go to a transportation - minimizing ( or equivalently price - maximizing ) location @xmath36 : @xmath37 ( notice that every @xmath36 yields the same minimal total expenditure to the customer living at @xmath12 ) . with the previous notation , the optimal pricing problem amounts to @xmath38 where @xmath39 is the functional @xmath40 by the definition of @xmath13 and @xmath41 , one has @xmath42 for all @xmath36 , hence the profit functional can be rewritten as : @xmath43 in order to obtain the existence of a solution to the optimization problem we reformulate the problem by using the variable @xmath44 instead of @xmath3 ; the advantage is that @xmath45 is searched among @xmath7-concave functions , while @xmath3 does not have special properties , and this will enable us to obtain the extra compactness necessary to prove the existence result . [ cconc ] a function @xmath46 is called @xmath7-concave if there exists a function @xmath47 ( that without loss of generality can be assumed upper semicontinuous ) such that @xmath48 for every @xmath7-concave function @xmath45 the @xmath7-transform @xmath49 is defined by @xmath50 and the @xmath7-superdifferential @xmath51 is given by @xmath52 the previous definition expresses that @xmath7-concave functions are functions that can be written as pointwise infima of functions of the form @xmath53 for some @xmath47 . the analogy with concave functions ( as infima of affine functions ) and the parallel between the @xmath7-transform and the more familiar legendre - fenchel transform then should be clear to the reader . in a similar way , the notion of @xmath7-superdifferential generalizes the notion of superdifferential for a concave function , and one can actually characterize @xmath7-superdifferentials in terms of the so - called @xmath7-cyclical monotonicity property that is analogous to the usual cyclical monotonicity . let us remark as a first example that if @xmath7 is a distance , then @xmath7-concave functions are exactly @xmath54-lipschitz functions , and in this case one can take @xmath55 in ( [ cconcave ] ) . the case of strictly convex costs , and in particular the quadratic cost , will be treated in subsection [ strconvex ] . [ unifcont ] every @xmath7-concave function is uniformly continuous and its continuity modulus is bounded by the continuity modulus of the cost function @xmath7 on @xmath27 . take a @xmath7-concave function @xmath45 and two points @xmath56 . by the definition of @xmath7-concavity , for a suitable upper semicontinuous function @xmath57 we have @xmath58 where @xmath59 is a suitable point in @xmath0 . then we have @xmath60 interchanging the role of @xmath61 and @xmath62 we deduce @xmath63 which concludes the proof . [ vc ] if @xmath64 is a sequence of @xmath7-concave functions converging uniformly to @xmath45 , then @xmath65 converge uniformly to @xmath49 and @xmath45 is @xmath7-concave . as a consequence , for every @xmath2 we have @xmath66 since @xmath65 are @xmath7-concave , by lemma [ unifcont ] it is enough to show that @xmath65 converge to @xmath49 pointwise on @xmath0 . fix @xmath35 and let @xmath67 be such that @xmath68 since @xmath0 is compact , a subsequence of @xmath69 converges to some @xmath2 , so that @xmath70 vice versa , if @xmath2 is such that @xmath71 , we have @xmath72 , so that @xmath73 the fact that @xmath45 is @xmath7-concave follows in an analogous way . for the last assertion , fixed @xmath2 take @xmath74 such that @xmath75 reaches on @xmath76 its minimal value @xmath77 . by definition of @xmath78 we have @xmath79 and we may assume that @xmath80 in @xmath0 . by the first part of the lemma we may pass to the limit and deduce that @xmath81 which gives @xmath82 and @xmath83 as required . we reformulate now problem by considering the functional @xmath84\,df(x)\ ] ] on the admissible class @xmath85 where @xmath86 by lemma [ unifcont ] the class @xmath87 is compact for the uniform convergence , and by lemma [ vc ] the optimization problem @xmath88 admits a solution @xmath89 . we can now come back to the initial problem and deduce that it admits an optimal solution @xmath90 . indeed , if @xmath91 , then @xmath92 ( since for any function @xmath57 we have @xmath93 ) and @xmath94 . moreover , it is easy to check that @xmath95 for every @xmath96 . thus , @xmath97 actually solves since for every @xmath96 one has @xmath98\,df(x)=i(v_p)\le i(v_{opt})=f(p_{opt}).\ ] ] so far , we have been in a rather abstract framework and it is time now to look at some special cases where the problem takes a more tractable form . we consider here the particular case when the cost function @xmath6 is given by a distance @xmath99 on @xmath0 ; we shall see that in this situation the solution @xmath90 above can be recovered in an explicit way . we denote by @xmath100 the class of all lipschitz functions in @xmath0 for the distance @xmath101 whose lipschitz constant does not exceed 1 . [ lipone ] in the case @xmath102 the optimal solution is given by @xmath103 we first notice that in this case the class of @xmath7-concave functions coincides with the class @xmath100 . moreover , as we have seen in the reduction from problem to problem , we may limit ourselves to consider only functions which are of the form @xmath104 where @xmath45 is @xmath7-concave . in our case this allows us to limit the class of admissible @xmath3 to @xmath100 . due to the tie - breaking rule it is easy to see that for @xmath105 it is @xmath106 , which gives to the cost functional @xmath39 the simpler form @xmath107 maximizing the previous expression in the class of functions in @xmath100 which are bounded by @xmath16 provides the solution @xmath108 as required . we notice that in the case @xmath102 above the optimal pricing pattern @xmath90 does not depend on the distribution @xmath29 of customers . note also the explicit formula for the optimal price : @xmath109 when @xmath0 is a subset of the euclidean space @xmath110 , then theorem [ lipone ] in particular applies to the concave case where @xmath111 with @xmath112 $ ] since such costs are in fact metrics . we consider now the case @xmath113 where @xmath114 is some open bounded subset of the euclidean space @xmath110 and @xmath115 where @xmath116 is a nonnegative smooth and strictly convex function . in this framework , a @xmath7-concave function @xmath45 can be represented as : @xmath117 by the smoothness of @xmath116 , the compactness of @xmath0 and lemma [ unifcont ] ensure that @xmath45 is lipschitz continuous on @xmath0 hence lebesgue a.e . differentiable on @xmath114 by rademacher s theorem . for every point @xmath118 of differentiability of @xmath45 and every @xmath119 , it is easy to check that from ( [ envcconv ] ) one has : @xmath120 and since @xmath116 is strictly convex this can be rewritten as : @xmath121 where @xmath122 stands for the legendre transform of @xmath116 . this proves that for every @xmath7-concave function @xmath45 , @xmath123 is in fact single - valued on a set of full lebesgue measure . now further assuming that @xmath29 is absolutely continuous with respect to the lebesgue measure on @xmath114 , we can rewrite the profit functional in a more familiar form : @xmath124\,df = \int_\omega\left[v+h^(\nabla v)-\nabla v\cdot\nabla h^(\nabla v)\right]\,df.\ ] ] if we further restrict our attention to the quadratic case , namely @xmath125 and @xmath114 is convex , it is easy to see that @xmath45 is @xmath7-concave on @xmath126 if and only if the function @xmath127 defined by @xmath128 is convex and satisfies @xmath129 of course the constraint @xmath130 translates into @xmath131 with @xmath132 . putting everything together , we then see that @xmath45 solves ( [ maxi ] ) if and only if @xmath133 and @xmath127 solves the following : @xmath134\,df\ ] ] and @xmath135 problems of the calculus of variations subject to a convexity constraint with a very similar structure as ( [ calcvarcvex ] ) arise in the monopoly pricing model of rochet and chon ( @xcite ) . note also that by strict convexity , ( [ calcvarcvex ] ) possesses a unique solution . we now consider problem ( [ calcvarcvex ] ) in the special unidimensional case where @xmath136 , @xmath137 and @xmath138 ( which corresponds to the price bound @xmath139 ) . the problem amounts to maximize @xmath140 among convex , nondecreasing and @xmath54-lipschitz functions @xmath127 . it is obvious that one necessarily has @xmath141 at the optimum , which setting @xmath142 and integrating by parts enables us to write @xmath143\,dx\ ] ] and the previous integral has to be minimized among nondecreasing functions @xmath144 taking values in @xmath145 $ ] . by a straightforward computation , the infimum is attained for @xmath146 , so that integrating we find @xmath147 and then by @xmath148 . summarizing , we have obtained @xmath149\\ -x^2/2+x-1/4 & \mbox { if } & x\in [ 1/2,1 ] . \end{array}\right.\ ] ] finally , the optimal price is obtained by the formula @xmath150 which simply yields here @xmath151 . in this subsection , we investigate the particular case where @xmath253 , the closure of a bounded open convex subset of @xmath8 and the cost is the euclidean distance @xmath254 . as before we assume that @xmath21 is an open subset of @xmath0 and @xmath255 . as already noticed , in this case , the @xmath7-concave functions are simply the @xmath54-lipschitz ones . as for the @xmath191-concave ones , it is easy to see that @xmath127 is @xmath191-concave if and only if it is @xmath54-lipschitz on @xmath0 and @xmath256 now , let @xmath257 be a point of differentiability of @xmath127 and let @xmath258 ( so that @xmath259 ) be such that @xmath260 ( i.e. @xmath252 ) , then one has @xmath261 so that @xmath262 ) holds a.e . on @xmath21 . in particular @xmath127 is an a.e . solution of the eikonal equation @xmath263 on @xmath21 . let @xmath257 be a point of differentiability of @xmath127 , @xmath252 and @xmath264 , then with the fact that @xmath127 is @xmath54-lipschitz , it is easy to check that @xmath265 , for every @xmath266 $ ] ( i.e. @xmath127 grows at the maximal rate @xmath54 on the segment @xmath267 $ ] ) . in particular , choosing @xmath268 $ ] such that @xmath269 yields : @xmath270 by density , this inequality actually holds for all @xmath257 , and the converse inequality follows immediately from ( [ repwp]).we thus have proved that if @xmath127 is @xmath191-concave then @xmath271 it is well - known ( see @xcite ) that ( [ repwpb ] ) implies that @xmath127 is a _ viscosity _ solution of the eikonal equation on @xmath21 . now , conversely , assume that @xmath127 is @xmath54-lipschitz on @xmath0 and a viscosity solution of the eikonal equation on @xmath21 and define @xmath272 then @xmath273 on @xmath189 ( in particular on @xmath274 ) and by the same argument as above @xmath57 is a viscosity solution of the eikonal equation on @xmath21 . a standard comparison argument ( e.g. theorem 2.7 in @xcite ) yields @xmath273 on @xmath21 so that @xmath127 is @xmath191-concave . this proves that the set of @xmath191 concave functions is : @xmath275 where the eikonal equation has to be understood in the viscosity sense . let us also remark that the condition @xmath276 a.e . in @xmath21 is in fact hidden in the definition of a viscosity solution ( equivalently in formula ( [ repwpb ] ) ) . getting back to our optimization problem ( [ pbmeenw ] ) , it is natural to introduce for every @xmath2 and @xmath277 ( the unit sphere of @xmath8 ) the quantity : @xmath278 for @xmath234 , we then have for a.e . @xmath2 @xmath279 assuming that @xmath29 is absolutely continuous with respect to the lebesgue measure on @xmath114 and defining @xmath207 by ( [ defwp ] ) , for @xmath280 , the profit functional @xmath281 is then given by : @xmath282 which has to be maximized over @xmath233 defined by ( [ formw ] ) . now , our aim is to transform the previous problem in terms of the values of @xmath127 on @xmath274 only . of course , because of ( [ repwpb ] ) , the behavior of @xmath127 on @xmath21 is fully determined by its trace on @xmath274 . in order to treat the behavior on @xmath189 , we need the following result . let @xmath234 and define @xmath283 then @xmath284 and @xmath285 . obviously @xmath284 and @xmath286 on @xmath189 hence the integrand in the definition of @xmath281 is larger on @xmath189 for @xmath57 than for @xmath127 ( recall that @xmath287 ) . if @xmath257 is such that @xmath244 , then the same conclusion holds now , if @xmath257 is such that @xmath288 , then we write @xmath289 with @xmath290 , if @xmath291 then @xmath292 and if @xmath293 then @xmath294 . since @xmath295 , in both cases we then have @xmath296 which proves that @xmath297 . in particular , @xmath298 a.e . on @xmath299 which proves the desired result . let @xmath280 and let @xmath300 be the trace of @xmath127 on @xmath274 , thanks to the previous lemma we may assume that @xmath301 on @xmath189 so that : @xmath302 because of ( [ repwpb ] ) , the second term only depends on @xmath300 , and the first one is monotone in @xmath127 hence for a given @xmath300 ( @xmath54-lipschitz and smaller than @xmath207 ) it is maximized by the largest @xmath54-lipschitz function on @xmath189 which has @xmath300 as trace on @xmath21 and is below @xmath207 i.e. simply @xmath303 since the previous formula also holds for @xmath257 by ( [ repwpb ] ) , we define for every @xmath54-lipschitz function @xmath300 on @xmath274 such that @xmath304 the state equation @xmath305 the profit maximization ( [ pbmeenw ] ) can thus be reformulated as the following nonstandard optimal control problem where the control is the price @xmath300 on the interface @xmath274 : @xmath306 where the class of admissible boundary controls @xmath307 consists of all @xmath54-lipschitz functions @xmath300 on @xmath274 such that @xmath308 and the state equation is ( [ state ] ) . for example if @xmath0 is the unit ball of @xmath8 and @xmath189 its boundary , then the maximization problem ( [ pbmeenw ] ) becomes maximizing : @xmath309 in the set of viscosity solutions of the eikonal equation @xmath263 on the unit ball . note that this is a highly nonconvex variational problem , which as previously may be reformulated as maximizing @xmath310 among @xmath54-lipschitz functions @xmath300 on @xmath274 such that @xmath304 . in the one dimensional case , the eikonal equation has a very simple structure which makes problem ( [ controlform ] ) much simpler . in particular , if @xmath274 is finite then the maximization of ( [ controlform ] ) reduces to a finite dimensional problem , since the control in this case is simply given by the values of @xmath127 on the finite set @xmath274 . for instance let us take @xmath311 $ ] , @xmath312 with @xmath313 . for simplicity let us also assume that @xmath16 is constant on @xmath21 and that @xmath29 is a probability that does not charge points . then the solutions of ( [ controlform ] ) only depend on the two scalars @xmath314 and @xmath315 subject to the constraints : @xmath316 for such a control @xmath317 the function @xmath318 has the following w - like shape : the function @xmath319 is in this case @xmath320\alpha,\beta[$ and $ \nu=1$,}\\ \beta - x&\mbox{if $ x\in]\alpha,\beta[$ and $ \nu=-1$,}\\ 0&\mbox{otherwise , } \end{array}\right.\ ] ] and the corresponding profit can be explicitly computed as a function of @xmath317 : @xmath321 where @xmath322 defining @xmath39 the cumulative function of @xmath29 ( i.e. @xmath323)$ ] ) , solving ( [ controlform ] ) then amounts to maximize : @xmath324 subject to the constraints ( [ contdf ] ) . for example , if @xmath325 , @xmath326 ( i.e. the price @xmath1 has to be chosen only at the boundary of @xmath0 ) and @xmath29 is uniform , then there is a unique optimal strategy given by @xmath327 . in what follows , @xmath0 will be some compact metric space ( the economic region ) , @xmath21 is some open subset of @xmath0 ( the subregion where prices are fixed ) and @xmath16 is a nonnegative l.s.c . function defined on @xmath21 ( @xmath16 is the fixed price system in @xmath21 ) . we are also given a transportation cost function @xmath7 assumed to be continuous and nonnegative on @xmath27 and such that @xmath28 for all @xmath2 . finally , @xmath29 is a nonnegative radon measure on @xmath0 of positive mass that models the repartition of customers in @xmath0 . we set @xmath152 , this ( compact ) subregion being the one where prices have to be determined by the agent . the unknown of the problem is the pricing pattern @xmath3 in the following class : @xmath153 once a price @xmath3 has been fixed by the agent , consumers living at @xmath12 purchase the good so as to minimize their total expenditure i.e. price plus commuting cost , which leads to define , analogously to what done in section [ firstpb ] , @xmath154 by our l.s.c . and compactness assumptions @xmath32 is a nonempty compact subset of @xmath0 and moreover the graph of @xmath13 is compact as the argmin of some l.s.c . function on @xmath27 . note that @xmath32 is the set of locations where consumers living at @xmath12 rationally choose to purchase the good . if @xmath22 then all the profit generated by the consumers of @xmath12 goes to the runner of region @xmath21 . we thus define : @xmath155 when @xmath156 , the agent makes no profit on consumers of @xmath12 ; when @xmath157 , we assume as _ tie - breaking rule _ that the consumers living at @xmath12 go to a transportation - minimizing ( or equivalently to a price - maximizing ) location @xmath36 : @xmath158 ( notice that every @xmath36 yields the same minimal expenditure to @xmath12 ) . with the previous notations , we see that the optimal pricing problem amounts to the maximization problem @xmath159 where @xmath160 by the definition of @xmath13 and @xmath41 , one has @xmath42 for all @xmath36 , hence the profit functional can be rewritten as : @xmath161 defining for all @xmath162 @xmath163 we may then rewrite in a more synthetical way the profit as @xmath164 [ rk1 ] so far , we have not assumed that @xmath3 has to be nonnegative , in fact this constraint is unnecessary since it will directly follow from the maximization problem ( [ pbmeenp ] ) . indeed if @xmath96 then @xmath165 is also in @xmath166 and @xmath167 . if @xmath168 on @xmath169 , this claim is obvious . we may then assume that @xmath170 . let @xmath157 be such that @xmath171 and let @xmath172 be such that @xmath173 ; we have @xmath174 and since @xmath175 this yields @xmath176 which implies @xmath177 , @xmath178 and @xmath179 . we then have @xmath180 to prove the existence of a maximizer in ( [ pbmeenp ] ) , we reformulate the problem in terms of @xmath41 rather than on the price @xmath3 which a priori does not have special properties . to do that , it is convenient to use some notions of generalized concavity that are natural in our context ( as well as in the monge - kantorovich theory ) . before introducing formal definitions let us remark that for @xmath181 , one can rewrite @xmath41 ( defined by ( [ defvptp2 ] ) ) as : @xmath182 ( where @xmath183 denotes the minimum of the two real numbers @xmath184 and @xmath185 ) with @xmath186 from the previous formula we see that @xmath187 can be represented as the pointwise infimum of a family of functions @xmath188 where the parameter @xmath4 takes its values in @xmath189 . this suggests the following definition . a function @xmath190 is called @xmath191-concave if there exists a function @xmath192 bounded from above such that @xmath193 if @xmath127 is @xmath191-concave there exists a kind of minimal representation ( as for the usual legendre - fenchel transform ) of @xmath127 in the form ( [ defccon ] ) . indeed , using the @xmath7-transform ( see definition [ cconc ] ) @xmath194 one has @xmath195 indeed , on the one hand , the definition of @xmath196 yields @xmath197 for every @xmath198 , hence : @xmath199 on the other hand , using the representation ( [ defccon ] ) yields @xmath200 on @xmath189 hence @xmath201 analogously to what was done in section [ firstpb ] , for every @xmath191-concave function @xmath127 , the @xmath191-superdifferential of @xmath127 at @xmath2 ( denoted @xmath202 ) is defined by @xmath203 since @xmath0 and @xmath189 are compact and @xmath191-concave functions and their @xmath7-transforms are continuous , it is easy to see that for every @xmath191-concave function @xmath127 and every @xmath2 , @xmath202 is a nonempty compact subset of @xmath189 and that @xmath204 is compact . the aim of this subsection is to reformulate the maximization problem ( [ pbmeenp ] ) in terms of @xmath205 only . let @xmath181 be nonnegative ( which is not restrictive in view of remark [ rk1 ] ) and write @xmath206 with @xmath205 and @xmath207 defined by ( [ defwp ] ) . then , let us define @xmath208 as already noticed , since @xmath205 is @xmath191-concave we have @xmath209 now let us define @xmath210 by construction @xmath211 hence @xmath212 . the next proposition expresses that the profit is improved when one replaces @xmath3 by @xmath213 . this allows us to restrict the analysis to prices that are ( up to a minus sign ) a @xmath7-transform on the free region @xmath189 and will enable us to express the problem in terms of @xmath127 only . more precisely , we have the following . [ reformul ] let @xmath96 , @xmath214 and let @xmath213 and @xmath127 be defined as above . then one has @xmath215 which imply @xmath216 and @xmath217 we already know that @xmath212 . using subsection [ gc ] we know that @xmath218 on @xmath189 . since @xmath214 and @xmath219 we have @xmath220 which proves @xmath221 and ( [ prop1re ] ) . now , fix @xmath162 and @xmath172 ; then @xmath222 which , since @xmath223 , implies that @xmath224 so that @xmath225 which proves ( [ prop2re ] ) . the previous argument also proves that @xmath226 . the fact that @xmath227 and ( [ prop4re ] ) are obvious . thanks to the fact that @xmath228 , the integrand in @xmath229 is nonnegative and thanks to ( [ prop2re ] ) we obtain @xmath230 on @xmath169 . by ( [ prop3re ] ) we then have : @xmath231 finally , thanks to ( [ prop3re ] ) and ( [ prop4re ] ) , @xmath229 can be rewritten as a function of @xmath127 only as in ( [ paw ] ) . proposition [ reformul ] thus enables us to reformulate the initial problem ( [ pbmeenp ] ) as : @xmath232 where @xmath233 is the set of all @xmath191-concave functions . more precisely , if @xmath127 solves ( [ pbmeenw ] ) then @xmath213 defined by @xmath210 solves ( [ pbmeenp ] ) . notice that if @xmath234 then one has for every @xmath235 latexmath:[\[\label{equic } that @xmath233 is an equicontinuous family . with the reformulation ( [ pbmeenw ] ) ( and the equicontinuity estimate ( [ equic ] ) ) at hand , we easily deduce the following existence result . [ exist ] problem ( [ pbmeenw ] ) admits at least one solution ( hence so does problem ( [ pbmeenp ] ) ) . let @xmath237 be some maximizing sequence of ( [ pbmeenw ] ) ; without loss of generality we may assume that the integrand in the definition of @xmath238 is nonnegative ( see remark [ rk1 ] ) and that @xmath239 . by ( [ equic ] ) we deduce that @xmath240 is uniformly bounded and equicontinuous . thanks to ascoli - arzela s theorem , passing to a subsequence if necessary we may assume that @xmath241 converges uniformly to some @xmath127 which is easily seen to be @xmath191-concave too . to prove that @xmath127 solves ( [ pbmeenw ] ) , we first use fatou s lemma : @xmath242 it is therefore enough to prove that for every @xmath2 @xmath243 if @xmath244 the right - hand side vanishes and , since @xmath245 uniformly , we have @xmath246 for @xmath247 large enough , so that the left - hand side vanishes too . assume now that @xmath248 , and let @xmath249 be such that @xmath250 passing to a subsequence if necessary we may assume that @xmath251 converges to some @xmath252 , hence ( [ scsx ] ) holds . in this section we propose some further developments of the optimization problems above that could be investigated . it is not our goal to enter into the details , which could possibly be treated in a future paper . the model problems considered in the previous sections could also be used to describe a two ( or more ) players game , where each player operates only on its own region and considers the prices on the other regions as fixed . more precisely , assume that @xmath328 where @xmath329 and @xmath330 are two compact sets with no interior point in common ( although this is not essential for what follows ) . on @xmath329 and @xmath330 two agents ( for instance the central governments of two different countries ) operate and initially two price functions @xmath15 and @xmath331 are present on @xmath329 and @xmath330 respectively . at a first step the agent that operates on @xmath329 modifies its price on @xmath329 considering @xmath332 fixed on @xmath330 and maximizes its income choosing an optimal price function @xmath333 ; then the agent that operates on @xmath330 plays its move considering @xmath333 fixed on @xmath329 and maximizing its income through the choice of an optimal price function @xmath334 . the game continues in this way then providing price functions @xmath335 and @xmath336 defined on @xmath329 and @xmath330 respectively . an interesting issue would be the study of the convergence of the sequences @xmath337 and @xmath338 to price strategies @xmath3 and @xmath339 that the two agents do not have the interest to modify any more . a related alternative is to consider the competitive problem between the agents operating on @xmath329 and @xmath330 as a two - persons game ( see for instance @xcite or @xcite ) , which is not zero - sum , where the strategy of each player is the pricing function on the region he controls . one has to be cautious in precisely defining the payoff functions when some customers are indifferent between being the good in @xmath329 or in @xmath330 . in such a case , one can for instance impose , as tie - breaking rule , that each customer shops in his own region , and for simplicity we assume @xmath340 . for respective price strategies @xmath3 ( prices on @xmath329 ) and @xmath339 ( prices on @xmath330 ) , define for all @xmath2 @xmath341 under our tie - breaking rule , the payoff functions for the two players are then given by @xmath342 defining admissible strategies as pairs of nonnegative and l.s.c . functions on @xmath329 and @xmath330 ( possibly also satisfying additional constraints ) , an interesting issue is then to find nash equilibria ( see for instance @xcite or @xcite ) for the payoffs @xmath343 , that is a pair of admissible strategies @xmath344 and @xmath345 such that @xmath346 this is a priori a complicated problem because kakutani s fixed - point theorem does not apply here because of the tie - breaking rule which induces discontinuities . of course , one can extend the framework to more than two - players , introduce mixed - strategies ... the analysis of spatial competition is an important issue in economics since hotelling s celebrated model @xcite and one may relate the equilibrium problem described above to this line of research . the study of the general nash problem is left for future research , but we give an elementary example where the solution is very simple and intuitive . let @xmath311 $ ] , @xmath347 $ ] , @xmath348 $ ] , @xmath349 and @xmath254 . as explained in section [ concave ] , given the strategy of the second ( respectively first ) player , the first ( resp . second ) one optimally choses a pricing function of the form @xmath350 ( resp . @xmath351 ) for @xmath352 ( resp . for @xmath353 ) . at a nash equilibrium one must have @xmath354 ( if @xmath355 then @xmath329 makes zero profit as well as @xmath330 makes zero profit if @xmath356 ) . finally , the common value @xmath354 has to be @xmath357 , since if for instance @xmath358 then @xmath330 can charge a slightly lower price @xmath359 at the border point @xmath360 then getting the whole demand and increasing his profit for @xmath361 small enough . in this simple case there is then a unique nash equilibrium @xmath362 and @xmath363 , no matter what the population distribution is . the equilibrium price is plotted in the next figure .	we consider an optimization problem in a given region @xmath0 where an agent has to decide the price @xmath1 of a product for every @xmath2 . the customers know the pricing pattern @xmath3 and may shop at any place @xmath4 , paying the cost @xmath5 and additionally a transportation cost @xmath6 for a given transportation cost function @xmath7 . we will study two models : the first one where the agent operates everywhere on @xmath0 and a second one where the agent operates only in a subregion . for both models we discuss the mathematical framework and we obtain an existence result for a pricing strategy which maximizes the total profit of the agent . we also present some particular cases where more detailed computations can be made , as the case of concave costs , the case of quadratic cost , and the onedimensional case . finally we discuss possible extensions and developments , as for instance the case of nash equilibria when more agents operate on the same market .
two of the outstanding issues for sne are the nucleosynthesis and the explosion mechanism . although sne have been known from the 1950 s to be the most important sources of heavy elements , there is little quantitative evidence for this . sn 1987a has in this respect been a unique source of information thanks to the possibility to obtain spectral information at very late epochs when the ejecta is transparent and the central regions observable . the nature of the explosion mechanism is still to a large extent unknown , and constraints from observations are badly needed . the main such constraints are provided by the hydrodynamic structure , e.g. , the extent of mixing , and the masses of the iron peak elements , in particular the radioactive isotopes ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? to model the spectra and light curve a complicated chain of thermalization processes from the gamma - rays and positrons in the kev mev range , to the observed optical and ir photons has to be understood . the details of this are discussed in a series of papers , @xcite . see also @xcite and @xcite . here we only summarize the main points . after a couple of days the main energy input to the sn ejecta comes from radioactive decay , first followed by . later than @xmath0 days takes over , while at very late epochs , @xmath1 days decay becomes dominant . table 1 summarizes the different decay routes , the emitted particle ( gamma - rays or positrons ) , the exponential decay time scale and the approximate epoch when the various decays dominated the energy input to sn 1987a . .main radioactive decays in sn 1987a [ cols=">,<,<,<,>,^",options="header " , ] to estimate the masses of the three most abundant radioactive isotopes formed in the explosion , , and , one can use several approaches , each with its special advantages and disadvantages ( fransson & kozma , in preparation ) . the best known , and for the mass most reliable , is the bolometric light curve . as shown by e.g. , @xcite , this yields a mass of @xmath2 . for , or rather , , which dominates the light curve between 1100 1800 days , this is not so straightforward , and a simple approach similar to that of , yielded a mass at least twice as large as other methods , and much higher than could be produced by the nucleosynthesis models @xcite . this was shown by @xcite to be a result of time dependent effects in the light curve . because of the decreasing density and ionization , the recombination time scale ( @xmath3 ) as well as the cooling time scale , become longer than the expansion time scale , @xmath4 . the gas is therefore not able to recombine and cool at the same rate as the radioactive input takes place , and only a fraction of the radioactive input is released as radiation . the difference is released later as recombination radiation , giving a higher luminosity than the instantaneous input . the bolometric light curve therefore gives an overestimate of the radioactive mass if a simple steady state calculation is employed . in figure [ fig3 ] we show a time dependent calculation of the bolometric light curve , and also showing the various individual radioactivity contributions . when we compare the bolometric luminosity with the total energy input one sees that between @xmath5 days and @xmath6 days the bolometric luminosity is up to a factor three higher than the input . at very late time most of the energy input is in the iron rich zones by positrons from , making time dependent effects less important . the most serious problem of using the bolometric light curve at late times is the large bolometric correction necessary because nearly all of the emission emerges in the unobserved far - ir . in figure [ fig4 ] we nevertheless show the time - dependent bolometric light curve . this represents an extension of the modeling in fransson & kozma ( 1993 ) to later epochs , and with updated physics . the freeze - out increases in this case the luminosity by a factor 2 3 at epochs later than 1200 days . the observations come from ctio data by @xcite , and eso data by @xcite . while the agreement with the ctio data is good , the agreement is considerably worse with the eso data during the late epochs . the discrepancy between the eso and ctio data probably gives a good estimate of the systematic errors in the observations , and it is clear that a different approach has to be taken for estimating the mass of , and also to get a more accurate estimate of the mass . co@xmath7 , m@xmath8co@xmath9 , and m@xmath10ti@xmath11 . data are from @xcite ( dots ) and from suntzeff ( 1992 ) ( triangles ) . the dotted line gives the steady state light curve , while the solid shows the same with time dependent effects included.,width=302 ] the second approach involves modeling of the line emission from the ejecta . an example of this was shown in figure [ fig1 ] , where we show light curves for three different values of the mass . . although the model reproduces the observations very well during the whole evolution , as a measure of the @xmath12ti mass the light curve is clearly not very sensitive . nearly all of the excitation is due to gamma - rays absorbed by the envelope , and the line is therefore strongly affected by the freeze - out . most of the emission at @xmath13 days is in fact due to the delayed recombination from the epoch at @xmath14 days when @xmath15co dominated the energy input . a better alternative is to use lines from the metal rich core , in particular the fe i and fe ii lines . an important result found by chugai ( 1997 ) is that in order to reproduce the fe ii flux , positrons have to be trapped efficiently in the iron rich gas . pure coulomb collisions are not sufficient for this , but a weak magnetic field is enough to make the gyro radius of the positrons small enough . a severe problem is , however , that modeling of individual lines is very sensitive to atomic data , which in spite of dramatic improvements from the iron - project , is still a problem . in a similar attempt @xcite have tried to use the fir [ fe ii ] @xmath16 fine structure line in connection with observations by iso . unfortunately , only an upper limit to the flux was obtained , yielding an upper limit to the mass of @xmath17 . there is , however , a number of loop holes in this result , such as extra dust cooling , mixing with other heavy elements contributing to the cooling like silicon and sulphur , as well as uncertainties in the atomic data . a compromise between these two approaches is offered by broad band photometry , where photometry in the optical and near - ir is compared to synthetic photometry from spectral calculations . this method avoids the need to extrapolate into the far - ir for the bolometric light curve , as well as many of the uncertainties involved in the calculation of individual lines . rather than modeling individual lines , the broad band fluxes are mainly sensitive to the division between uv , optical and ir in the spectral distributions of the elements this is determined by the atomic configuration , i.e. the term diagram , and less by the exact path a recombining photon takes when cascading to the ground state . except for the fraction of the emission in the optical and ir which is absorbed by the dust , the photometry is insensitive to this component . the fraction absorbed by dust can be estimated from mid - ir observations , as well as from line profile shifts ( lucy 1991 ) . in figure [ fig5 ] we show the b and v band light curves together with observations . the reason for concentrating on these bands is that they are strongly dominated by lines from fe i and fe ii , and therefore most sensitive to the mass , in contrast to the r band which is dominated by , and therefore sensitive to the freeze out . observations are taken from suntzeff ( 1991 ) and hst observations from suntzeff ( 2001 ) . the models have @xmath18 of @xmath19co , @xmath20 of @xmath15co and three different values of the mass , @xmath21 . as can be seen , both the b and v bands are well reproduced by the observations throughout the whole evolution , and the currently best estimate of the mass is in the range @xmath22 . this method also yields the best estimate of the mass , @xmath23 . a more complete discussion of these results will be given in fransson & kozma ( 2001 ) . of @xmath12ti . all models have @xmath20 of @xmath15ni , width=377 ] sn 1987a is by far the best source of nucleosynthesis of any sn . reliable masses for both the most abundant elements , h , he , and o , as well as the the most abundant radioactive isotopes , , , and have been obtained . these provide some of the most important constraints on the progenitor and on the explosion itself .	the nucleosynthesis and production of radioactive elements in sn 1987a are reviewed . different methods for estimating the masses of , , and are discussed , and we conclude that broad band photometry in combination with time - dependent models for the light curve gives the most reliable estimates . supernovae , nucleosynthesis , radioactivity 97.60.bw , 26.30.+k
the trace distance is one of the most natural distance measures used in quantum - information theory @xcite . it is one of the main tools used in distinguishability theory and it is connected to the average success probability when distinguishing two states by a measurement @xcite . it is also related to quantum fidelity , which provides the measure of similarity of two quantum states @xcite . both quantities are particularly important in quantum cryptography , since the security of quantum protocols relies on an ability to measure the distance between two quantum states @xcite . the trace distance is also related to other properties of quantum states like the von neumann entropy and relative entropy @xcite . the main aim of this work is to provide a lower bound for the trace distance using measurable quantities . we use nonlinear functions of the form @xmath0 , where @xmath1 and @xmath2 are density matrices . for such forms there exist feasible schemes to measure them in an experiment without resorting to state tomography @xcite . we give a lower bound based on the superfidelity introduced recently in @xcite . proof of this bound gives an answer to the conjecture stated by mendonca _ _ in @xcite . let us denote by @xmath3 the space of density matrices acting on @xmath4-dimensional hilbert space @xmath5 . for two density matrices @xmath6 the trace distance is defined as @xmath7 in the particular case of pure states we can use bloch vectors @xmath8 and @xmath9 . one can see that the trace distance between such states is equal to half of the euclidean distance between the respective bloch vectors @xmath10 the trace distance can be bounded with the use of the fidelity @xcite @xmath11 where the fidelity is defined as @xmath12 ^ 2.\ ] ] the inequality ( [ eqn : fid - dtr - ineq ] ) shows that @xmath13 and @xmath14 are closely related indicators of distinguishability . the main result of this work is a lower bound for the trace distance , which we prove in the next section , @xmath15 where @xmath16 is called the _ superfidelity _ , and was introduced in @xcite . for @xmath17 it is defined as @xmath18 from the matrix analytic perspective the inequality ( [ eqn : main - ineq ] ) relates the trace norm on the space @xmath3 to the hilbert - schmidt scalar product on @xmath3 . for the sake of consistency we provide basic information about the superfidelity @xcite . the most interesting feature of superfidelity is that it provides an upper bound for quantum fidelity @xcite @xmath19 the superfidelity also has properties which make it useful for quantifying the distance between quantum states . in particular we have : ( 1 ) bounds : @xmath20 . ( 2 ) symmetry : @xmath21 . ( 3 ) unitary invariance : for any unitary operator @xmath22 , we have @xmath23 . ( 4 ) concavity : @xmath24 for any @xmath25 and @xmath26 $ ] . ( 5 ) supermultiplicativity : for @xmath27 we have @xmath28 note that the superfidelity shares properties 1 - 4 with the fidelity . however , in contrast to the fidelity , superfidelity is not multiplicative , but supermultiplicative . in @xcite the authors showed that @xmath16 is _ jointly _ concave in its two arguments , @xmath29 for @xmath26 $ ] . note that the property of joint concavity is obeyed by the square root of fidelity @xmath30 , but not by the fidelity ( [ eqn : fid ] ) . fidelity can be used to define the metric @xmath31 on the space @xmath3 as @xmath32 unfortunately the analog of the bures distance defined using the superfidelity @xmath33 is not a metric @xcite , but the quantity @xmath34 provides the metric on the space @xmath3 . finally one should note that the superfidelity is particularly convenient to use as the practical measure of similarity between quantum states . one of the main advantages of superfidelity is that it is possible to design feasible schemes to measure it in an experiment @xcite . also , from the computational point of view , calculation of the superfidelity is significantly less resource - consuming @xcite . the properties of superfidelity listed above suggest that it would be convenient to use it instead of fidelity to draw conclusions about the distinguishability of quantum states . this section show how this can be done by relating the superfidelity and trace distance . first we can observe that from the inequality @xmath35 and since we have ( [ eqn : fid - dtr - ineq ] ) we get that @xmath36 our main aim is to prove the following inequality , which provides tighter bound . [ conjecture - general ] for any @xmath6 we have @xmath37 or equivalently @xmath38 this inequality was first stated as a conjecture in @xcite , where it was verified numerically for small dimensions . clearly it is motivated by the lower bound for trace distance provided by the inequality ( [ eqn : fid - dtr - ineq ] ) . to prove the theorem [ conjecture - general ] we need the following lemma . for any @xmath6 let @xmath39 and @xmath40 be the projectors onto @xmath41 and @xmath42 respectively . we have the following inequalities @xmath43 _ proof . _ because of the similarity we will show only inequality ( [ ineq:1 ] ) . it is easy to prove inequalities ( [ ineq:2 ] ) , ( [ ineq:3 ] ) and ( [ ineq:4 ] ) in a similar manner . we subtract the right- from the left - side of ( [ ineq:1 ] ) to get @xmath44 because @xmath45 is positive semidefinite . @xmath46 _ proof of theorem [ conjecture - general ] . _ adding the inequalities ( [ ineq:1 ] ) and ( [ ineq:2 ] ) we get @xmath47 similarly , by adding inequalities ( [ ineq:3 ] ) and ( [ ineq:4 ] ) we get @xmath48 now we notice that , if two non - negative numbers are greater than the third one , then so is the geometric mean of the first two numbers . using this fact we combine ( [ eqn : l1 ] ) and ( [ eqn : l2 ] ) to get @xmath49 on the other hand we can rewrite the trace distance with the use of projectors @xmath50 and @xmath51 : @xmath52 now finally we can write @xmath53 which is equivalent to ( [ eqn : conjecture - general ] ) . @xmath46 it is natural to consider the relation between the lower bounds in ( [ eqn : fid - dtr - ineq ] ) and ( [ eqn : main - ineq ] ) . it is clear that bound ( [ eqn : main - ineq ] ) is better than ( [ eqn : fid - dtr - ineq ] ) whenever @xmath54 . in the one qubit case or if one of the states is pure the bound ( [ eqn : main - ineq ] ) is always better than ( [ eqn : fid - dtr - ineq ] ) . this follows from the equality between fidelity and superfidelity in these situations @xcite . on the other hand the inequality ( [ eqn : fid - dtr - ineq ] ) provides a better lower bound if the states @xmath55 and @xmath56 have orthogonal supports . in this case the fidelity between states vanishes , but the superfidelity is not necessarily equal to zero @xcite . to get some feeling about the difference between the bound given by fidelity and the present bound we will consider the following families of states . 1 . the family @xmath57 is defined as @xmath58 where @xmath59 is a pure state and in our case we take @xmath60 . the family @xmath61 is defined as @xmath62 where @xmath63 . the family @xmath64 is defined as : @xmath65 calculated for states ( [ eqn : ex - states ] ) and maximally mixed states as a function of the dimension @xmath4 and the parameter @xmath66 [ see eq . ( [ eqn : ex - states ] ) ] . ] calculated for states ( [ eqn : ex - states ] ) and ( [ eqn : ex - states - dim4 ] ) as a function of parameters @xmath66 and @xmath67 . ] in fig . [ fig : low - bound - diff ] we consider the family @xmath57 and calculate the difference @xmath68 . one can see that for small dimensions and states close to the pure state @xmath59 the superfidelity gives a much better approximation for the trace distance than the fidelity . for larger dimensions this is not the case , but nevertheless we can observe that still the superfidelity provides a better bound . this advantage is lost for states close to the maximally mixed state . a similar situation can be observed in fig . [ fig : low - bound - diff - dim4 ] where the difference between @xmath69 and @xmath70 ( @xmath71 ) is presented . for this particular family the bound ( [ eqn : main - ineq ] ) is better than ( [ eqn : fid - dtr - ineq ] ) , but the difference vanishes for states close to the maximally mixed state . calculated for states ( [ eqn : ex - states ] ) and ( [ eqn : ex - states - k3 ] ) ( @xmath72 ) as the function of parameters @xmath66 and @xmath73 . solid line is a border that separates two regions . for the parameters in lighter one the inequality ( [ eqn : fid - dtr - ineq ] ) provides better bound , while in darker region the inequality ( [ eqn : main - ineq ] ) is better . ] unfortunately , the bound ( [ eqn : main - ineq ] ) is not always tighter when compared with ( [ eqn : fid - dtr - ineq ] ) . figure [ fig : diff - levels ] shows the difference between @xmath74 and @xmath75 ( @xmath72 ) . as one can see for this family of states there are regions for which the bound ( [ eqn : main - ineq ] ) is better than ( [ eqn : fid - dtr - ineq ] ) but in this case for highly mixed states ( [ eqn : fid - dtr - ineq ] ) is better than ( [ eqn : main - ineq ] ) . we know that the probability of error for distinguishing two density matrices @xmath17 is expressed by the trace distance as @xcite @xmath76.\ ] ] using the inequality ( [ eqn : main - ineq ] ) we can write @xmath77 we have shown the relation between the superfidelity and trace distance , which is analogous to the relation with trace distance and fidelity . this shows that superfidelity can be used to conclude about the distinguishablity of states . our bound provides the relation between the trace distance and the overlap of two operators supplementary to inequality from ( * ? ? ? * th.1 ) . as such it provides neat mathematical tool which can be used in quantum information theory . authors would like to thank k. yczkowski , r. winiarczyk and p. gawron for many interesting , stimulating and inspiring discussions . we acknowledge the financial support by the polish ministry of science and higher education under the grant number n519 012 31/1957 .	we provide a bound for the trace distance between two quantum states . the lower bound is based on the superfidelity , which provides the upper bound on quantum fidelity . one of the advantages of the presented bound is that it can be estimated using a simple measurement procedure . we also compare this bound with the one provided in terms of fidelity .
in recent years , high angular resolution x - ray telescopes make it possible to detect x - ray sources with only a few counts . this is very different from the optical photometry . because of these low counts , the poisson processes in corresponding wavebands can not be approximated to gaussian distribution . therefore the statistics will be very different in some estimations and calculations than used before . in recent years bayesian method has gained many applications ( e.g. van dyk et al . 2001 and references therein ) since it has more advantages in low count cases than traditional statistics . hardness ratios are widely used in high energy astrophysics since faint sources with only limited counts can not give any satisfying spectral modeling . in x - ray detection , hardness ratios are normally used to show spectral properties roughly ( e.g. tennant et al . , 2001 ; sivakoff , sarazin & carlin 2004 ) . hardness ratios are usually defined as the ratio of counts in different wavebands ( @xmath1 ) or the ratio of the difference and sum of counts in two wavebands ( @xmath2 ) , @xmath3 and @xmath4 are counts in two wavebands @xmath5 and @xmath6 . on the other hand , for the spectra of x - ray sources , the photoelectric absorption , quantified with the hydrogen column density @xmath7 , can not be neglected . hydrogen column density contains many kinds of important information , such as the radial distance and the interstellar circumstance of the sources . for low count sources , hydrogen column density is hard to know since no reliable spectral fitting can be made . however interstellar absorption is energy dependent . consequently the information of hydrogen column density can be drawn from the hardness ratios . in this paper , we first give a new definition of hardness ratio and its estimation method , and then we discuss the procedure to estimate the hydrogen column density accordingly . we begin our discussion with the following problem : suppose that one experiment obtained two counts from two different poisson distributions , we need to : ( 1 ) estimate the ratio of the expectation values of the two poisson distributions , and ( 2 ) construct the confidence interval of the ratio . the expectation values of the poisson processes are just the @xmath0 parameters of the poisson distribution @xmath8 . therefore the above problem may be formulated as follows : suppose @xmath3 and @xmath4 are two counts corresponding to two different poisson processes @xmath5 and @xmath6 with their parameters as @xmath9 and @xmath10 respectively , @xmath11 and its confidence interval needs to be estimated . to solve this problem we first need to derive the distribution of @xmath0 parameter under certain counts , i.e. , the conditional distribution of @xmath9 and @xmath10 , as follows . @xmath12 first we assume , as a pragmatic convention , a uniform prior for the @xmath0 parameter . @xmath13 similarly , @xmath14 this continuous distribution is gamma distribution , as shown in fig . 1 . in addition , we use jeffreys prior ( @xmath15 ) , which may be more advantageous over the uniform prior commonly used , because the inferences derived from jeffreys prior are parameterization - invariant ( see kass & wasserman 1996 for detail of this prior ) . under this prior , the conditional distribution of @xmath0 is @xmath16 and @xmath17 to account for the background contamination , suppose that @xmath3 is the count corresponding to a poisson process @xmath5 with the addition of a poisson background process @xmath18 , the expectation value of the process @xmath18 is assumed to be known as @xmath19 . according to the properties of poisson processes , the sum of two poisson processes is also a poisson process with the parameter @xmath20 . so the probability of the total count @xmath21 is @xmath22 . apply the bayesian assumption and the uniform prior distribution assumption , we obtain the conditional distribution of @xmath9 , as follows . @xmath23 when @xmath19 is much smaller than @xmath3 , this result is same as equation(2 ) . also we can obtain the conditional distribution under the jeffreys prior distribution assumption , @xmath24 and it is same as equation(4 ) when @xmath19 is much smaller than @xmath3 . there are two different definitions of hardness ratio , @xmath25 and @xmath26 . in traditional method , the estimate of @xmath27 and @xmath28 are @xmath29 and @xmath30 respectively , and the errors are propagated under the gaussian distribution , i.e. , @xmath31 here we propose a method to estimate the hardness ratio based on the bayesian method . both the uniform prior and the jeffreys prior will be used . when @xmath19 is much smaller than @xmath3 , we use equation(2 ) ( under the uniform prior ) or equation(4 ) ( under the jeffreys prior ) to estimate @xmath25 and @xmath32 . first we assume the uniform prior of @xmath0 . for the conditional distribution function of @xmath11 , @xmath33 for the conditional probability density function of @xmath11 , @xmath34 this distribution is shown in fig . 2 when @xmath35 , @xmath36 . it is easy to verify that the distribution is normalized , @xmath37 for the hardness ratio @xmath28 , the probability distribution of this hardness ratio is given by , @xmath38 the conditional probability density function is given by , @xmath39 this distribution is shown in fig . 3 when @xmath35 , @xmath36 . it is easy to verify that the distribution is normalized , @xmath40 when @xmath35 , @xmath36.the solid line represents the distribution derived by our proposed method under the uniform prior , the doted line represents the distribution derived by our proposed method under the jeffreys prior , the dashed line represents the gaussian distribution derived by the traditional method . ] when @xmath35 , @xmath36.the solid line represents the distribution derived by our proposed method under the uniform prior , the doted line represents the distribution derived by our proposed method under the jeffreys prior , the dashed line represents the gaussian distribution derived by the traditional method . ] similarly , we obtain the conditional probability density function of @xmath27 and @xmath28 under the jeffreys prior assumption as follows , @xmath41 and @xmath42 using the result of @xmath43 and @xmath44 , we can estimate @xmath27 and @xmath28 . under the uniform prior assumption , since we have only one observation , we take the most probable value as the estimate of @xmath25 , denoted as @xmath45 . let @xmath46 we obtain , @xmath47 similarly , we obtain the most probable value as the estimate of hr : @xmath48 similarly , we get the most probable value of @xmath27 and @xmath28 under the jeffreys prior assumption , @xmath49 and @xmath50 the highest posterior density ( hpd ) interval is used to give the error bars . the hpd interval under the confidence level @xmath51 is the range of values which contain a fraction @xmath51 of the probability , and the probability density within this interval is always higher than that outside the interval . there are other point estimates and error estimates . for example , the mean value and the equal tailed interval . since the distributions of @xmath27 and @xmath28 are obtained , these alternative estimates can be easily derived . when @xmath19 can not be ignored , equation ( 6 ) or equation ( 7 ) can be used to estimate @xmath27 and @xmath28 . in this situation , it is difficult to give a simple analytic distribution function like equation ( 10 ) or equation ( 13 ) ; in this case , one can only use numerical integration to obtain the distribution of @xmath27 and @xmath28 , and then do the estimate . we use the monte carlo simulations to investigate the statistical properties of our result , and compare them with the traditional methods . first we set @xmath9 and @xmath10 . do poisson sampling for @xmath52 times and each time we get @xmath3 and @xmath4 respectively . each time we estimate @xmath27 and @xmath28 using two kinds of methods . finally we obtain that , for two methods , the mean square error of the point estimate , the coverage rate ( the percentage of times during which the confidence interval contains the real value ) , and the mean confidence interval . the simulations contain two cases : low counts and high counts . in case 1 , we first set @xmath53 and @xmath54 , then set @xmath55 and @xmath54 . in case 2 , we first set @xmath56 and @xmath57 , then set @xmath56 and @xmath58 . the confidence level in the simulations is 90% . the results of the simulations are shown in table 1 . .statistical properties of our method and traditional method . [ cols="^,^,^,^,^,^,^ " , ] [ park : tbl : coverage ] from the simulation results , we notice that our proposed method is more reliable than the traditional method when the counts are low . the reason is that the traditional method is based on using the gaussian distribution to approach the poisson distribution , which is not reliable when the counts are low . here we propose a method to estimate the hydrogen column density using data obtained with the _ chandra _ x - ray observatory . because of the high angular resolution of _ chandra _ , a positive detection of a point source during a survey observation only requires several counts . therefore our method will be more reliable when estimating the hardness ratio than the traditional method . the detail procedure of this application can be found in another paper ( wu et al . the basic idea is introduced as follows . the basic procedure consists of the following three steps : ( 1 ) calculate the relationship between the hardness ratios and @xmath7 values under certain spectral model ; ( 2 ) estimate hardness ratios according to observed counts in different wavebands ; and ( 3 ) interpolate the @xmath7 values and error intervals from hardness ratios . according to the most likely physical nature of the sources , we can assume a spectral model ( e.g. power law with photon index @xmath59 for typical x - ray binaries ) . then we can use _ pimms _ ( http://cxc.harvard.edu/toolkit/pimms.jsp ) tool to calculate the relationship between hydrogen column density and the hardness ratio . using pimms _ , one can get the count rate in certain energy band under a given x - ray spectrum and hydrogen column density . the count rate in the given energy band is just the @xmath0 parameter in this energy band ; this was our original movivation of defining the hardness ratios in terms of the @xmath0 parameter . the calculated @xmath7 ( hydrogen column density ) @xmath28 relationships are shown in fig . 7 for three different energy bands of @xmath5 ( 1 - 3 kev ) , @xmath6 ( 3 - 5 kev ) and @xmath60 ( 5 - 8 kev ) for a _ acis - i observation ( wu et al . 2006 ) respectively : @xmath61 , @xmath62 . from fig . 7 , we can see that @xmath63 or @xmath64 is more appropriate for @xmath65 or @xmath66 , respectively . having the value and error interval of @xmath28 , we can finally do linear interpolation on curves in fig . 7 to obtain the value and error interval of hydrogen column density . first we give the conditional probability distribution of @xmath0 parameter under certain counts in a poisson process using bayesian statistics . according to this result we derive the probability density function of two kinds of hardness ratios . we take the most probable values as the estimate of hardness ratios and the hpd intervals as the error intervals . then we use monte carlo simulations to investigate the statistical properties of our results , and find that our method is more reliable than the traditional method when the counts are low . finally we show how to estimate the hydrogen column density using hardness ratios . our method developed in this paper provides a way to estimate the hydrogen column density of sources which are too faint to do spectral fitting . however the spectral shape for these sources must be assumed _ a prior_. this method is especially convenient for a sample of faint sources with similar spectra . after this paper has been submitted initially on 06 - 03 - 29 , we noticed another submitted paper ( park et al . 2006 ) which discusses the same statistical problem as we have done in this paper . in that paper the authors also used the bayesian method to estimate the hardness ratio , and showed some applications on quiescent low - mass x - ray binaries , the evolution of a flare , etc , therefore justifying the wide range of applications of such a statistical problem . since the strict analytic solution of the hardness ratio distribution does not exist for general situations , the authors suggested methods by monte carlo and numerical integration to obtain the distribution in that paper . in our paper we find simple analytic solutions of the probability density functions of hardness ratios for the situations in which the background can be ignored . this will be useful and convenient for some applications , such as _ chandra _ data in which background can be ignored for hardness ratio estimation of point sources . finally , we note , under the advise of the referee , that in 1980s some studies have been done on the ratio of poisson means both from a frequentist standpoint ( james & roos 1980 ) and from a bayesian standpoint ( helene 1984 and prosper 1985 ) . in this paper we used the bayesian method under the uniform prior and the jeffreys prior , made extensive comparisons between this method and traditional method , aiming explicitly at applications in astrophysics . xie chen read the first draft carefully and gave many helpful suggestions , especially on english writing . we are particularly grateful to the referee for his pointing out relevant historical literature in other fields and suggesting using jeffreys prior . this study is supported in part by the special funds for major state basic research projects and by the national natural science foundation of china ( project no.10233030 , 10327301 and 10521001 ) . 99 kass , r. e. , & wasserman , l. , 1996 , j. amer . 91 , 1343 james , f. , & roos , m. , 1980 , nucl . b172 , 475 helene , o. , 1984 , nucl . instr . and meth . a228 , 120 prosper , h. b. , 1985 , nucl . instr . and meth . a241 , 236 park , t. , kashyap , v. l. , siemiginowska , a. , dyk , d. a. v. , zezas , a. , heinke , c. , & wargelin , b. j. , 2006 , submitted to apj ( arxiv : astro - ph/0606247 ) sivakoff , g. r. , sarazin , c. l. , & carlin , j. l. , 2004 , apj , 617 , 262 tennant , a. f. , wu , k. , ghosh , k. k. , kolodziejcazk , j. j. , & swartz , d. a. , 2001 , apj , 549 , l43 van dyk , d. a. , connors , a. , kashyap , v. l. , & siemiginowska , a. , 2001 , apj , 548 , 224 wu , j. f. , zhang , s. n. , lu , f. j. , & y. k. jin , 2006 , accepted for publication in chinese journal of astronomy and astrophysics ( arxiv : astro - ph/0606478 )	hardness ratios are commonly used in x - ray photometry to indicate spectral properties roughly . it is usually defined as the ratio of counts in two different wavebands . this definition , however , is problematic when the counts are very limited . here we instead define hardness ratio using the @xmath0 parameter of poisson processes , and develop an estimation method via bayesian statistics . our monte carlo simulations show the validity of our method . based on this new definition , we can estimate the hydrogen column density for the photoelectric absorption of x - ray spectra in the case of low counting statistics .
the cholesteric phase can be considered as a special case of the nematic phase @xcite . the long axes of the anisotropic molecules are on the average aligned parallel to each other within planes . the direction of this alignment rotates smoothly as one proceeds in a direction perpendicular to the parallel planes . such a phase ( the twisted uniaxial nematic phase ) will be called the uniaxial cholesteric phase . by analogy , the twisted biaxial nematic phase will be called the biaxial cholesteric phase . we note that in the case of the twisted biaxial nematic phase three twists around three perpendicular axes of biaxial symmetry should be considered simultaneously . there are many models of the uniaxial cholesteric phase where molecules are assumed to be uniaxial . in 1970 goossens @xcite showed within the extended maier - saupe model that the quadrupole interactions give rise to the twist . in 1976 straley @xcite presented a statistical - mechanical theory of the elastic constants and of the spontaneous twisting of a cholesteric . he expressed the pitch by means of the elastic constants . we note that he gave the geometric interpretation of the chiral term which is often present in microscopic theories of cholesterics . this term can approximately describe the minimum approach distance of a pair of threaded rods . we will derive similar terms in the case of biaxial molecules . in 1977 lin - liu _ @xcite presented a molecular theory of cholesteric liquid crystals and showed the general form of the chiral potential energy of interactions for uniaxial molecules . they discussed different types of the temperature dependence of the pitch . real molecules forming liquid crystalline phases are never uniaxial and it is more realistic to assume less symmetric biaxial molecules . some properties of the cholesteric phase formed by biaxial molecules were investigated in the past . in 1974 priest and lubensky @xcite found the order of the biaxial order parameter and predicted a fluctuation instability . van der meer and vertogen @xcite discussed the dependence of the pitch on the biaxial order parameter . they also derived expressions for the temperature dependence of the pitch and the elastic constants @xcite in the case of the uniaxial cholesteric phases . finally , in 1992 evans @xcite presented a hard body model for chiral nematic liquid crystals . the density functional theory was used , and molecules were represented by a hard convex twisted ellipsoidal core , with and without an encircling isotropic square well . the pitch was found to be density and temperature independent with values in the visible region of the spectrum . long range potential softness could account for the increase of the pitch with decreasing temperature . the phase behaviour of liquid crystalline mixtures has been studied , using a number of theoretical methods . in 1980 sivardiere @xcite introduced the ising - like model and obtained a large variety of phase diagrams . @xcite considered the maier - saupe model and gave a cataloque of allowed diagrams for mixtures of nematogens . as far as the cholesteric mixtures are concerned , the theoretical and experimental works describe usually uniaxial cholesteric phases composed of uniaxial molecules . in 1971 @xcite studied the helical twisting power in nematic - cholesteric mixtures as a function of concentration . it was found that , as the concentration of the cholesteric material increases , the twisting power of the mixtures increases lineary up to a certain concentration , beyound which it increases more slowly , taking a maximum value at some specific concentration . then , in the high - concentration region , the twisting power falls . we note that typically the additivity rule is satisfied as described in sec . [ sec : corner ] . in 1977 lin - liu _ @xcite presented a molecular theory of binary cholesteric mixtures . they considered uniaxial molecules and derived a formula showing how the pitch depends on temperature and concentration . under certain conditions the formula reduces at fixed @xmath0 to a simple quadratic rational fraction in the composition . the forming of the cholesteric phase is closely connected with chirality . some authors tried to establish quantitative relationships between molecular properties and measurable properties that result from chiral molecular structures . @xcite proposed an intrinsic molecular chirality tensor based only on nuclear position . the chirality tensor gives rise to two universal chirality indices , the first givig information about absolute chirality , and the second about anisotropy of the chirality . recently harris _ et al . _ @xcite showed that any chiral measure of a geometric object is a pseudoscalar and must involve three - point correlations that only come into play when the molecule has at least four atoms . in general , a molecule is characterized by an infinite set of chiral parameters . however , one can also consider chirality of interactions @xcite and we will use this approach . our aim is to describe the uniaxial and biaxial cholesteric phases that consist of biaxial or uniaxial molecules . we would like to derive the microscopic expressions for the elastic constants and the pitch . we will investigate the inversion in the context of chirality . our paper is organized as follows . in sec . [ sec : pheno ] we present a phenomenological continuum theory of cholesteric liquid crystals . in sec . [ sec : micro ] we describe a statistical theory of cholesteric phases that consist of rigid biaxial molecules , and in sec . [ sec : elastic ] we derive general expressions for the elastic constants and the cholesteric pitch . a condition of stability is obtained that concerns both nematics and cholesterics . exemplary calculations are presented in sec . [ sec : corner ] , were the corner potential energy is applied . by means of symmetry considerations we will identyfy main chiral terms for biaxial molecules . in the limit of uniaxial molecules we will recover the results from @xcite . in sec . [ sec : summary ] we summarize the results of this work . in this section we will describe a uniform phase from a phenomenological point of view @xcite . we assume that at every point @xmath1 inside a considered phase we can define three orthonormal vectors @xmath2 reflecting orientational properties of this phase . in the case of the biaxial phase they determine directions of its two - fold axes of symmetry . the vectors @xmath3 create the local frame which can be expressed by means of a space - fixed reference frame @xmath4 as @xmath5 where repeated indices imply summation . the completely ordered uniform phase is described by @xmath6 . let us call @xmath7 the free energy due to the distortion of the local frame @xmath3 . a general form of its density @xmath8 was derived in @xcite in the case of small distortions . when a considered phase has a @xmath9 symmetry group ( the biaxial cholesteric phase ) we get @xmath10 where @xmath11 , @xmath12 , @xmath13 are the elastic constants , @xmath14 the terms with @xmath15 give 3 bulk terms , the terms with @xmath16 give 3 surface terms , the terms with @xmath17 give 12 bulk and 3 surface terms of the form @xmath18 the total numbers of bulk and surface terms are 15 and 6 , respectively . if a considered phase possesses a @xmath19 symmetry group ( the uniaxial cholesteric phase ) , the number of the elastic constants is smaller , because some constants from the previous case become dependent or zero . let the @xmath20 axis be oriented along the axis of symmetry . then the distortion free - energy density has the form of the frank expression @xmath21^{2 } + { \frac { 1}{2 } } k_{3 } [ \vec{n } \times ( \nabla \times \vec{n } ) ] ^{2 } \nonumber \\ & + & { \frac { 1}{2 } } k_{4}\nabla \cdot [ ( \vec{n } \cdot \nabla ) \vec{n } - \vec{n } ( \nabla \cdot \vec{n } ) ] \nonumber \\ & + & { \frac { 1}{2 } } k_{5}\nabla \cdot [ ( \vec{n } \cdot \nabla ) \vec{n } + \vec{n } ( \nabla \cdot \vec{n})],\end{aligned}\ ] ] where the relations among the nonzero elastic constants are @xmath22 therefore in the case of the uniaxial phase we have 4 bulk ( from @xmath23 to @xmath24 ) and 2 surface terms ( @xmath25 and @xmath26 ) . in the continuum theory of uniaxial nematic liquid crystals three basic types of deformations i.e. splay , twist and bend , appear , which extract from the distortion free energy terms with @xmath27 , @xmath28 and @xmath24 , respectively . thus each constant @xmath29 must be positive ; if not , the undistorted nematic conformation would not correspond to a minimum of the free energy @xmath7 . in @xcite 18 basic deformations proper for the continuum theory of biaxial nematics were given . they were divided into five groups and connected with relevant elastic constants : 3 twists ( for @xmath30 ) , 6 splays and bends ( for @xmath31 ) , 3 modified twists ( for @xmath16 ) and two groups of 3 double twists ( for @xmath32 and for @xmath33 ) . inside the formulas for deformations a parameter @xmath34 was used ( @xmath35 was a certain length ) . small @xmath34 meant a small deformation and a conformation close to the uniform one @xmath36 . the vectors of the local frame were expanded into a power series with respect to @xmath34 : @xmath37 it appeared that the most important in eq . ( [ eq : l = l0+el1 ] ) were terms linear in @xmath34 . they were sufficient to calculate the distortion free energy up to the second order in @xmath34 , and to calculate the elastic constants of biaxial nematic liquid crystals . we add that for @xmath15 one should use the same deformations as for @xmath30 because the relevant terms have different dependence on the parameter @xmath34 and they do not mix . in this section we focus on the microscopic analysis of cholesteric liquid crystals . let us consider a dilute gas of @xmath38 molecules contained in a volume @xmath39 , at the temperature @xmath40 ( in kelvine ) . let it be a binary mixture of @xmath41 molecules @xmath42 and @xmath43 molecules @xmath44 , where @xmath45 . we assume that molecules are rigid blocks with three translational and three rotational degrees of freedom . the state of a molecule @xmath46 is described by a vector of the position @xmath47 and the orientation @xmath48 , where @xmath49 , @xmath50 and @xmath51 are the three euler angles . on the other hand , one can use the set of the three orthonormal vectors @xmath52 . in a space - fixed reference frame @xmath4 we can express them as @xmath53 let @xmath54 , @xmath55 , @xmath56 and @xmath57 denote the mass of a molecule @xmath58 ( @xmath59 or @xmath44 ) and the three moments of inertia , respectively . apart from that , we denote @xmath60 , @xmath61 are the three euler angles for @xmath62 . in fact , @xmath63 is not used , because @xmath64 and @xmath65 will appear inside the rotation matrix elements @xmath66 . we assume that molecules interact via two - body short - range forces which depend on the distance between molecules and their orientations ; @xmath67 gives the potential energy of interactions ( @xmath68 ) . now we would like to discuss in detail symmetries of the energy @xmath69 and molecules , because it is crucial for the forming of the cholesteric phase . we note that the euler angles always enter any formulas via the standard rotation matrix elements @xmath70 . there is a problem how to incorporate the inversion @xmath71 into this formalism because the inversion can not be expressed by means of rotations . however , one can try to simulate the inversion . we start from the vector @xmath62 . after the inversion operation it gives @xmath72 . let us denote the euler angles after the inversion by @xmath73 . @xmath74 we used an unknown function @xmath75 of the angle @xmath63 because the third euler angle is not determined before and after the inversion . note that the relation ( [ eq : inversion1 ] ) is often generalized to the form @xcite @xmath76 the relation ( [ eq : inversion2 ] ) in general is not true because it cancels the matrix elements @xmath70 with @xmath77 odd . on the other hand such elements have to be used as described in @xcite . let us rewrite the relation ( [ eq : inversion1 ] ) in the form @xmath78 thus , for the matrix elements @xmath66 the inversion can be connected with the rotation @xmath79 and the @xmath80 symmetry group . this means that there is no difference between @xmath19 and @xmath81 within the considered formalism . we postulate that the invariance with respect to the inversion means the invariance with respect to the three separate rotations @xmath82 , @xmath83 and @xmath84 and one should replace @xmath85 with @xmath86 , @xmath87 and @xmath88 , respectively . this connects @xmath71 with the @xmath9 symmetry group . this means that there is no difference between @xmath9 and @xmath89 within the considered formalism . it is known that @xmath90 so only two rotations have to be used . now we are in a position to discuss all symmetries of the potential energy @xmath69 . 1 . _ translational invariance : _ it is satisfied because @xmath91 depends on @xmath92 . _ rotational invariance : _ @xmath91 should not depend on a choice of a reference frame . it means that for any rotation @xmath93 @xmath94 we add that generally it does not have to be satisfied because the interactions between two molecules can be modified by the presence of others molecules , especially in an ordered phase . invariance with respect to the permutation of identical molecules : _ @xmath95 4 . _ invariance with respect to the symmetry operations of molecules : _ for biaxial molecules we apply operations from the @xmath96 symmetry group . for the molecule @xmath58 @xmath97 @xmath98 and for the molecule @xmath99 @xmath100 @xmath101 if the molecule @xmath58 is uniaxial , we should add operations from the @xmath81 symmetry group . for any @xmath102 @xmath103 5 . _ invariance with respect to complex conjugation : _ @xmath91 should be a real function . _ chirality of the interactions : _ the interactions between molecules are _ nonchiral _ when @xmath104 where @xmath105 denotes the euler angles @xmath93 after the inversion . in the opposite case the interactions are chiral and this leads to the forming of the cholesteric phase . the microscopic free energy of the binary mixture has the form @xcite @xmath106 - 1 \right\ } \nonumber \\ & & - { \frac { 1}{2 } } \sum_{i , j = a , b } \int { d(1)}{d(2 ) } g_{i}(1 ) g_{j}(2 ) f_{12}^{ij } , \end{aligned}\ ] ] where @xmath107 are the one - particle distribution functions with the normalizations @xmath108 @xmath109 , @xmath110 are the mayer functions , @xmath111 , and @xmath112 our set of state variables consists of @xmath40 , @xmath39 , @xmath113 , and @xmath114 . the free energy ( [ eq : betafmix ] ) consists of the ideal terms ( with @xmath115 ) and the excess terms directly related to intermolecular forces . the ideal terms are that of the ideal gas . the expression ( [ eq : betafmix ] ) was derived systematically for binary mixtures from the bogoliubov - born - green - kirkwood - yvon hierarchy equations in the thermodynamic limit @xmath116 @xcite . the two - particle distribution functions were expressed in terms of the one - particle distribution functions and the two - particle correlation functions of the simple form @xmath117 . this assumption guarantees the proper limit of the unary system . the equilibrium distributions @xmath118 minimizing the free energy ( [ eq : betafmix ] ) satisfy @xmath119 - \sum_{j = a , b } \int { d(2 ) } g_{j}(2 ) f_{12}^{ij } = \mbox{const } .\ ] ] in the homogeneous phase the distribution function @xmath118 do not depend on the position of a molecule and @xmath120 . in order to obtain @xmath121 one should solve the equations ( [ eq : lng - sum_int ] ) together with ( [ eq : normagi ] ) . in order to define the microscopic distortion free - energy density @xmath122 one should also identify the homogeneous free - energy density @xmath123 . note that the terms with @xmath115 in ( [ eq : betafmix ] ) are local : the integrands involve the distribution function for a single point only . other terms couples the distributions at neighbouring points . the dependence of the free energy on the spatial variations of the ordering will be found by expanding this terms in the gradients of @xmath124 @xcite . substituting the taylor expansion of @xmath125 into ( [ eq : betafmix ] ) we get @xmath126,\ ] ] where @xmath127 - 1 \right\ } \nonumber\\ & & \mbox { } - { \frac { 1}{2 } } \sum_{i , j = a , b } \int { dr_{1}}{dr_{2 } } { d\vec{u } } g_{i}(\vec{r},r_{1 } ) g_{j}(\vec{r},r_{2 } ) f_{12}^{ij } , \label{def : f0 } \\ \beta f_{d}(\vec{r } ) & = & - { \frac { 1}{2 } } \sum_{i , j = a , b } \int { dr_{1}}{dr_{2 } } { d\vec{u } } g_{i}(\vec{r},r_{1 } ) [ ( \vec{u } \cdot \nabla)g_{j}(\vec{r},r_{2 } ) ] f_{12}^{ij } \nonumber\\ & & \mbox { } + { \frac { 1}{4 } } \sum_{i , j = a , b } \int { dr_{1}}{dr_{2 } } { d\vec{u } } [ ( \vec{u}\cdot\nabla)g_{i}(\vec{r},r_{1 } ) ] \nonumber\\ & & \times [ ( \vec{u}\cdot\nabla)g_{j}(\vec{r},r_{2 } ) ] f_{12}^{ij}. \label{def : fd}\end{aligned}\ ] ] note that an integration by parts has been used to combine the second order terms and the surface terms have been neglected thanks to the thermodynamic limit . the definition ( [ def : f0 ] ) is equivalent to that by poniewierski and stecki @xcite . this is a well - founded assumption if we also assume slow variations of the vectors @xmath3 . we will also restrict the one - particle distribution function @xmath118 to the class of @xmath121 functions . this method was succesfully used in the past @xcite . as we expect , for the homogeneous phase @xmath122 becomes equal to zero . it was shown in @xcite that in the case of the homogeneous biaxial nematic phase composed of biaxial molecules , the one - particle distribution function @xmath121 depends on four arguments : @xmath128 we postulate that the distribution of the distorted phase @xmath129 can be written as @xmath130 where the reference frame @xmath4 is replaced with the local frame @xmath2 . now we are in a position to substitute the basic deformations into the microscopic distortion free - energy density ( [ def : fd ] ) and to the phenomenological distortion free - energy density ( [ eq : d2 ] ) and ( [ eq : d8 ] ) . as a result of the comparison we get the microscopic formulas for the elastic constants . to make them more compact we write @xmath131 microscopic expressions for the chiral elastic constants of the biaxial cholesteric phase are as follows : @xmath132 microscopic expressions for the nonchiral elastic constants are the following . the first group is , @xmath133 the second group is , @xmath134 the third group is , @xmath135 the fourth group is , @xmath136 , \\ \beta k_{2233 } & = & { \frac{1}{4 } } \int { dr_{1}}{dr_{2}}{d\vec{u } } u_{y}u_{z } \nonumber \\ & & \times \sum_{i , j = a , b } f_{12}^{ij } [ -u_{1y}^{i}(u_{2z}^{j}-w_{2x}^{j})-(u_{1z}^{i}-w_{1x}^{i } ) u_{2y}^{j } ] , \\ \beta k_{1133 } & = & { \frac{1}{4 } } \int { dr_{1}}{dr_{2}}{d\vec{u } } u_{x}u_{z}\sum_{i , j = a , b } f_{12}^{ij } [ -u_{1y}^{i } w_{2y}^{j}-w_{1y}^{i}u_{2y}^{j}].\end{aligned}\ ] ] the fifth group is , @xmath137 in the case of the uniaxial cholesteric phase the expression for the chiral elastic constant is @xmath138 the expressions for the nonchiral elastic constants are as follows : @xmath139 let us show how the elastic constants determine the state of cholesteric liquid crystals . it is very important that , when we look for the physical state of the liquid crystal in the thermodynamic limit , we should remove all surface terms from the phenomenological free - energy densities . the reason is that in the thermodynamic limit all surface terms are negligible . we used the surface terms to derive the expressions for the elastic constants to assure the consistency because the surface terms are present in the hidden form in the microscopic free - energy density . the equilibrium distortion of the uniaxial cholesteric phase is a pure twist @xmath140\ ] ] where @xmath141 , @xmath142 is the cholesteric pitch and @xmath143 . note that the sign of @xmath23 distinguishes between right- and left - handed helices . for the biaxial cholesteric liquid crystals the equilibrium distortion is composed of three twists with respect to the orthogonal axes . in the limit of small distortions the vectors of the local frame have the form @xmath144 where @xmath145 , @xmath146 the distortion free - energy density at the minimum is @xmath147/2w^{2}.\end{aligned}\ ] ] it is important that a global minimum of @xmath122 exists only if the matrix @xmath148\ ] ] is positive definite . in the opposite case there is no stable biaxial cholesteric ( and nematic ) phase . taking the uniaxial phase limit , we get simpler conditions . we can say that there is a stable cholesteric or nematic phase only if @xmath149 as far as we know almost all theories and all real and computer experiments are in the agreement with the relation ( [ eq : k1<3k2 ] ) . the equality @xmath150 appears in some theories with hard molecules @xcite . the aim of this section is to express the elastic constants by means of the order parameters which can be measured in experiments . we will apply the corner potential energy of the interactions because in principle it allows detailed calculations without any additional approximations . on the other hand , it is quite realistic . the corner potential energy has the form @xmath151 , where @xmath152 depends on orientations @xmath153 , @xmath154 and @xmath65 . for @xmath152 one can write the general expansion proposed by blum and torruela @xcite . it involvs the 3-j wigner symbols and the standard rotation matrix elements . the same expression was used to describe the interactions of biaxial molecules in @xcite . in the case of biaxial molecules the lowest order terms of the expansion give @xmath155 \nonumber \\ & & \mbox { } + \sigma_{72}^{ij } [ ( \vec{n}_{1 } \cdot \vec{l}_{2 } ) ( \vec{n}_{1 } \times \vec{l}_{2 } ) \cdot \vec{\delta } - ( \vec{n}_{1 } \cdot \vec{m}_{2 } ) ( \vec{n}_{1 } \times \vec{m}_{2 } ) \cdot \vec{\delta } ] \nonumber \\ & & \mbox { } + \sigma_{8}^{ij } [ ( \vec{l}_{1 } \cdot \vec{l}_{2 } ) ( \vec{l}_{1 } \times \vec{l}_{2 } ) \cdot \vec{\delta } + ( \vec{m}_{1 } \cdot \vec{m}_{2 } ) ( \vec{m}_{1 } \times \vec{m}_{2 } ) \cdot \vec{\delta } \nonumber \\ & & \mbox { } - ( \vec{l}_{1 } \cdot \vec{m}_{2 } ) ( \vec{l}_{1 } \times \vec{m}_{2 } ) \cdot \vec{\delta } - ( \vec{m}_{1 } \cdot \vec{l}_{2 } ) ( \vec{m}_{1 } \times \vec{l}_{2 } ) \cdot \vec{\delta } ] .\end{aligned}\ ] ] there are 13 molecular parameters that determine the main features of the interactions between biaxial molecules . we would like to add some comments to this long expression . the terms with @xmath156 , @xmath157 , @xmath158 , and @xmath159 describe the nonchiral interactions of uniaxial molecules . the term with @xmath160 describes chiral part of their interactions . a similar term was used in the past @xcite . the terms with @xmath156 , @xmath157 , @xmath161 , @xmath162 , @xmath163 , @xmath164 , @xmath165 , and @xmath166 describe the nonchiral interactions of biaxial molecules @xcite . the additional terms with @xmath160 , @xmath167 , @xmath168 , and @xmath169 allow to decribe the chiral interactions of biaxial molecules . as far as we know they were not presented in the literature . note that the terms with @xmath62 couple the translational and rotational degrees of freedom . in order to set the proper values of the parameters @xmath170 one should investigated the equipotential surfaces of the potential energy . the excluded volume method described in @xcite can be helpful . molecules @xmath58 are described by @xmath171 the interactions between molecles @xmath58 and @xmath99 we describe by @xmath172 the parameters @xmath170 are the molecular constants that in our model determine the ordering of the phase , the value of the elastic constants and the twist of the phase . as far as the functional dependence of the potential energy on the @xmath173 is concerned , we have many possibilities and we will mention two of them , together with a function @xmath174 defined as @xmath175.\ ] ] let @xmath176 denotes a depth of the potential energy ( we assume for simplicity that it is the same for both types of molecules ) . 1 . the _ soft - core potential energy _ @xmath177 @xmath178 2 . the _ square - well potential energy _ @xmath180.\ ] ] let us define dimensionless functions @xmath181 with the normalization @xmath182 in order to describe systems with biaxial symmetry it is convenient to introduce the complete set of basic functions ( invariants ) proper for that kind of symmetry . they are closely connected with the rotation matrix elements and have the form @xmath183 where @xmath184 is a nonnegative integer , @xmath185 are even . if @xmath184 is even , then @xmath186 and @xmath187 . if @xmath184 is odd , then @xmath188 and @xmath189 . the invariants are real and orthogonal functions @xmath190 the properties of the invariants were described in @xcite . it was shown that all @xmath191 can be expressed by means of @xmath192 , @xmath193 , @xmath194 , and @xmath195 . let us show the most important invariants with @xmath196 : @xmath197/2 , \nonumber\\ f_{02}^{(2 ) } & = & [ -1+(\vec{n}\cdot\vec{e}_{z})^2 + 2 ( \vec{l}\cdot\vec{e}_{z})^2]\sqrt{3}/2 , \nonumber\\ f_{20}^{(2 ) } & = & [ -1+(\vec{n}\cdot\vec{e}_{z})^2 + 2 ( \vec{n}\cdot\vec{e}_{x})^2]\sqrt{3}/2 , \nonumber\\ f_{22}^{(2 ) } & = & [ -3+(\vec{n}\cdot\vec{e}_{z})^2 + 2 ( \vec{l}\cdot\vec{e}_{z})^2 + 2 ( \vec{n}\cdot\vec{e}_{x})^2 + 4 ( \vec{l}\cdot\vec{e}_{x})^2]/2.\end{aligned}\ ] ] for the first time all four invariants ( with different normalization ) appeared in the paper by straley @xcite . let us define the order parameters for our system as @xmath198 note that for the completely ordered biaxial phase we have @xmath199 , thus the most important are @xmath200 and @xmath201 ( @xmath202 and @xmath39 in the straley notation ) . let us rewrite ( [ eq : lng - sum_int ] ) for the one - particle distribution function of the uniform phase . it is identical to one in the theory with nonchiral interactions . we define @xmath203 @xmath204 the kernel @xmath205 and the distribution functions @xmath206 can be expressed in terms of the invariants @xmath207 @xmath208 note that for @xmath152 given by ( [ def : sigma ] ) both sums ( [ sumk ] ) and ( [ sumlnf ] ) are _ finite _ series . thus the distribution function @xmath206 is fully described by 35 coefficients @xmath209 . we calculate the coefficients @xmath209 from the equations @xmath210 @xmath211 now we move to the elastic constants . we can expand the distribution function @xmath206 in an infinite series with respect to invariants @xmath212 all elastic constants can be written as _ finite _ sums of the form @xmath213 @xmath214 where @xmath215 @xmath216 the coefficients @xmath217 , @xmath218 , and @xmath219 are polynomials in @xmath220 that can be calculated analytically . for simplicity in our calculations we took into account only the dependence on the order parameters @xmath221 and @xmath222 . we assumed the square - well potential energy of interactions with @xmath223 . note that @xmath224 determines the length scale , whereas @xmath176 determines the energy scale . the elastic constants will be expressed in @xmath225 and @xmath226 , temperatures will be given in @xmath227 . we assumed that molecules are similar to ellipsoids with three different axes @xmath228 . we used @xmath229 and our set of the molecular parameters is @xmath230 in our system , on decreasing the temperature we meet the first - order transition to the uniaxial cholesteric phase at @xmath231 and the second - order transition to the biaxial cholesteric phase at @xmath232 . the temperature dependence of the order parameters @xmath221 and @xmath222 is presented in fig . the order parameters were used to calculate the temperature dependence of the elastic constants . it appeared that the factors @xmath233 and @xmath234 introduce very strong variations with temperature and it was more comfortable to present the elastic constant divided by those factors as shown in figs . [ fig2 ] and [ fig3 ] . the corresponding twist parameters @xmath34 and @xmath235 are shown in fig . [ fig4 ] . in the uniaxial cholesteric phase there is one chiral constant ( @xmath23 ) and three nonchiral ones ( @xmath27 , @xmath28 , and @xmath24 ) . the twist parameter @xmath34 describing the pure twist is decreasing with temperature . for some potential energies one can obtain also the increasing temperature dependence , more often present in cholesterics . we have also the accidental equality @xmath236 caused by the omission of the order parameters @xmath237 with @xmath184 greater than 2 . typically , for rodlike molecules we get the inequalities @xmath238 . in the biaxial cholesteric phase we have three chiral constants ( @xmath239 , @xmath240 , and @xmath241 ) and 12 nonchiral ones . on decreasing the temperature , @xmath23 splits into @xmath242 and @xmath241 becomes nonzero . @xmath27 splits into @xmath243 , @xmath24 splits into @xmath244 , @xmath28 splits into @xmath245 , @xmath246 becomes a new independent constant . other new independent constant are four positive @xmath247 , @xmath248 , @xmath249 , @xmath250 and one negative @xmath251 . they are about two orders of magnitude smaller then the previous ones . the twist parameters @xmath235 are quickly increasing with the decreasing temperature and satisfy inequalities @xmath252 . we mixed the prolate molecules interacting via the soft core potential with @xmath253 . the set of molecular parameters is @xmath254 where @xmath224 denotes some molecular length and @xmath255 . the elastic constants will be expressed in @xmath256 and @xmath257 , the temperature in @xmath258 . for simplicity reasons we assume that only one phase is present . it corresponds to the stable solution with the minimum free energy . the possibility of phase separation will not be investigated . we checked the stability against perturbations of nematic symmetry and the stability against deformations of the phase . the molecules @xmath42 and @xmath44 in the unary phase form the uniaxial cholesteric and nematic phase , respectively . the transition temperatures of unary systems from the isotropic to the uniaxial cholesteric or nematic phase are @xmath259 and @xmath260 . the temperature dependence of the order parameters @xmath261 and the elastic constants for unary systems are shown in figs . [ fig5 ] and [ fig6 ] . the pitch of the system of molecules @xmath42 is a falling concave function of the temperature . in the considered range of temperatures the isotropic or uniaxial cholesteric phase was present in the mixture . the transition temperature @xmath262 of the binary system from the isotropic to the uniaxial cholesteric phase can be described approximately as @xmath263 we performed a detailed analysis of the mixture at the temperature @xmath264 where the uniaxial cholesteric phase was present for all concentrations . the order parameters and the elastic constants are continous and concave functions of the concentration @xmath265 . the dependence of the elastic constants on the concentration is plotted in fig . [ fig7 ] . the twist of the mixture is approximately described by the equation @xmath266 where @xmath267 are related to the unary systems at the same temperature . the additivity rule ( [ eq : q = sumxq ] ) is satisfied for most mixtures between cholesterol derivatives @xcite but there are some exceptions @xcite . note that in general @xmath267 may be positive or negative . in this paper we developed the statistical theory for the uniaxial and biaxial cholesteric phases . rigid molecules interacting via two - body short - range forces were assumed . we derived the microscopic formulas for the elastic constants and the pitch of the phase . in order to calculate the values of the elastic constants one needs the one - particle distribution function and the potential energy of the molecular interactions . it is showed that the elastic constants determine the twist of the phase . the obtained stability condition is to our knowledge not to be found in the literature . in the case of the uniaxial cholesteric or nematic phase it has the simple form @xmath268 . it seems that it is satisfied for all known substances . our theory was applied to the unary and binary systems of molecules similar to ellipsoids with three different axes . thanks to the corner potential energy the elastic constants were expressed as a finite series of the order parameters . apart from this the role of the temperature is more transparent : the function @xmath269 determines the order parameters ; the functions @xmath270 and @xmath271 are closely related to the chiral and nonchiral elastic constants , respectively . the temperature dependence of the order parameters , the elastic constants and the twists in uniaxial and biaxial cholesteric phase was obtained . the results concernig the nonchiral elastic constants @xmath17 are the same as in the case of nematic phases described in @xcite . we focus on the new results on the chiral elastic constants @xmath15 . during the transition from a uniaxial to a biaxial cholesteric phase @xmath23 splits into @xmath239 and @xmath240 , whereas @xmath241 becomes nonzero . below a certain temperature they satisfy inequalities @xmath272 . the corresponding twist parameters satisfy similar relations @xmath273 . note that those relations depend on the type of molecular interactions and do not have to be always satisfied . in the binary mixture it appeared that the transition temperatures and the twist of the mixture approximately satisfy the simple additivity rules . our theory is a starting point for further researches . it is desirable to go beyound a low density limit where the mayer function is replaced with a better approximation of the direct correlation function @xmath274 . the theory can be generalized to the case of more component liquid crystalline mixtures . we are also waiting for the experimental data on the biaxial nematic and cholesteric phases and the elastic constants as , according to kini and chandrasekhar , experiments are feasible @xcite . the experiments can verify the theoretical predictions and provide new problems . this research was supported by the state committee for scientific research ( kbn ) , grant no . 2p 03b05415 . temperature dependence of the chiral elastic constants @xmath275 . in the uniaxial cholesteric phase ( between @xmath231 and @xmath232 ) we have @xmath276 and @xmath277 . in the biaxial cholesteric phase ( below @xmath232 ) @xmath23 splits into two constants and @xmath241 becomes nonzero . ] temperature dependence of the nonchiral elastic constants @xmath278 . in the uniaxial cholesteric phase ( between @xmath231 and @xmath232 ) we have @xmath279 equal to @xmath280 and @xmath281 . we have also @xmath282 . in the biaxial cholesteric phase ( below @xmath232 ) @xmath27 , @xmath28 and @xmath24 split into two constants , @xmath246 becomes an independent constant , five constants become nonzero ( @xmath247 , @xmath251 , @xmath248 , @xmath249 and @xmath250 ) . ] temperature dependence of the twist parameters @xmath34 and @xmath235 . in the uniaxial cholesteric phase ( between @xmath231 and @xmath232 ) we have @xmath283 . in the biaxial cholesteric phase ( below @xmath232 ) all @xmath235 are positive and @xmath273 . ]	a statistical theory of cholesteric liquid crystals composed of short rigid biaxial molecules is presented . it is derived in the thermodynamic limit at a small density and a small twist . the uniaxial ( biaxial ) cholesteric phase is regarded as a distorted form of the uniaxial ( biaxial ) nematic phase . the chirality of the interactions and the implementation of the inversion to the rotation matrix elements are discussed in detail . general microscopic expressions for the elastic constants are derived . the expressions involve the one - particle distribution function and the potential energy of two - body short - range interactions . it is shown that the elastic constants determine the twist of the phase . the stability condition for the cholesteric and nematic phases is presented . the theory is used to study unary and binary systems . the temperature and concentration dependence of the order parameters , the elastic constants and the twist of the phase are obtained . the possibility of phase separation is not investigated . _ key words : _ liquid crystals ; cholesterics ; elastic constants ; mixtures .
the laws of quantum mechanics allow for quantum computers@xcite , which are known to be significantly more powerful than classical computers . in a quantum computer , information is stored in quantum bits ( qubits ) , rather than classical bits@xcite . a single qubit represents a zero or one and is a two - level system with two energy levels which can be used to store and process the information@xcite . an example of a two - level system , frequently used in quantum optics , is composed of the ground and excited states of an atom . semiconductor qds can be modeled as a two - level system with one exciton in the excited state@xcite . semiconductor qds embedded inside a cavity has been a subject of intense research as a promising candidate for quantum computation and information tasks , as well as a source of single photon@xcite . recently , experiments on the cavity embedded qds have been reported to show different spectral features of the mollow - triplet fluorescence spectrum@xcite . in particular , a dot coupled to the acoustic phonon bath in the super - ohmic regime has shown modified spectral features as a function of the phonon bath temperature . more precisely , the triplet sideband is observed to show a systematic spectral sideband broadening for both resonant and off - resonant cases . this problem was studies experimentally@xcite and anlyzed theoretically@xcite in terms of the usual born - markov approximation . however , the pure dephasing process a 3d phonon bath is in the form a super - ohmic independent boson model ( ibm ) , which is known to be highly non - markovian@xcite , and these results have been studied and discussed in terms of usual born - markov approximation@xcite . the system correlation function , for non - markovian interactions ( for e.g. nuclear spins@xcite , phonons@xcite ) , decays with a typical time scale given by the correlation time @xmath0 which never dies - off to zero . in other words , the correlation time is non - zero and can be larger than the system decay time @xmath1 , which is the signature of a strongly history - dependent non - markovian interaction . furthermore , the equation of motion for correlation function has an additional term known as irrelevant part , which is non - zero for non - markovian interactions . this additional term vanishes for a markov approximation , when the well - known qrt can be applied to find the system correlation@xcite . recent theories [ for e.g. ref . ] discussing fluorescence spectrum in solid - state systems rely on a history - independent markov process and apply the usual qrt , giving rise to an exponential decay of the system correlation . often , physical processes in solid - state systems@xcite are highly non - markovian ( history - dependent ) , and so the resulting spectrum using the qrt can no longer be used to describe their spectral properties . in this paper , we analyze the effect of pure dephasing due to a 3d acoustic phonon bath on the fluorescence spectrum of a cavity coupled to a semiconductor qd . the associated emission spectrum can be directly related to a correlation function for system observables , which we evaluate beyond the markov approximation using a nakajima - zwanzig gme@xcite . assuming a large band - width , coupling to the radiation modes of the cavity can be treated within born - markov approximation , which is relevant to cavity quantum electrodynamics ( cavity - qed ) experiments@xcite . the dot is represented by a two - level system with an exciton in the excited state coupled to a phonon reservoir , and can be modeled with the usual ibm@xcite . the resultant fluorescence spectrum has three components due to dressing of the levels by a pump laser@xcite , and we project the system into a dressed state basis which allows us to characterize the three components of triplet separately@xcite . we have solved the dot - phonon coupling using an exact approach beyond born approximation , and the exact result coincides with that of obtained within born approximation . phonon coupling gives rise to a frequency dependent frequent - shift and dephasing , which bring non - lorentzian features in the fluorescence spectrum . frequency - shift and dephasing due to phonon interaction are strongly temperature - dependent and vanish for the lower temperatures . we found that in the dressed - state basis levels of interest are coupled asymmetrically to the phonons and have vanishing dephasing and frequency - shift for a resonant dot - laser frequency . we have also observed that the sideband undergoes a distinct narrowing and becomes asymmetric with increasing temperature , which is explained using the dressed - states energy levels . this paper is organized as follows : in sec . [ sec : fluorescence ] , we start with discussing the setup and establish the formula for resonance fluorescence spectrum of a general two - level system . in sec . [ sec : model ] , we introduce the model hamiltonian for a driven cavity - qed two - level system interacting with a phonon bath . in sec . [ sec : gme ] , we discuss and derive the exact form of nakajima - zwanzig gme for the dynamics of the reduced density matrix and correlation function . we obtain the expressions for the lineshape functions in both markovian and non - markovian regimes . in sec . [ sec : results ] , we present our results and discuss the plots in different parametric regimes . in sec . [ sec : conclusions ] , we conclude with a discussion and summary of the results . other technical details are discussed in the appendixes . we start with the model hamiltonian of a general two - level system interacting with radiation modes of the electromagnetic field , which can be written as system ( @xmath2 ) , field ( @xmath3 ) , and interaction ( @xmath4 ) in terms of the standard jaynes - cummings model within a rotating - wave approximation ( rwa ) : @xmath5 where @xmath6 and @xmath7 are the raising and lowering operators between excited state @xmath8 and ground state @xmath9 in the hilbert space of the system and @xmath10 , @xmath11 are the annihilation and creation operators in the hilbert space of a set of electromagnetic modes coupled to the system . coupling to the radiation modes is given by a coupling constant @xmath12 to the mode of frequency @xmath13 . for simplicity of the units , we have also used @xmath14 . here , we adopt the well - known theory of gedanken spectrum analyzer given in ref . [ ] , and assume that the radiation field emitted by the system is detected by a two - level atom ( detector ) with a transition frequency @xmath15 , which is prepared in its ground state @xmath16 initially , see fig . [ fig : tls : rwa ] . is driven by a laser pump frequency @xmath17 . a phonon bath , shown by the green lines , is coupled to the excited state @xmath18 of the qd . ( b ) in the rotated frame with respect to the pump laser , where @xmath19 is detuning of the transition ( cavity , probe ) frequency from the laser . here , role of the spectrum analyzer is to admit frequencies which are resonant with the transition frequency of the detector . the photonic density of states of the cavity is described by a lorentzian spectrum , and @xmath20 is the central frequency of the cavity mode.,scaledwidth=50.0% ] the hamiltonian for the detector is given by @xmath21 and coupling between detector atom @xmath22 and the radiation field , included in the hamiltonian @xmath23 , is given by the hamiltonian @xmath24 where @xmath25 is the coupling of detector to a field mode @xmath26 . therefore , the hamiltonian for entire system including system and detector , as well as coupling to the radiation modes can be written as : @xmath27 according to wiener - khintchine theorem , fluorescence spectrum , in the stationary regime and in the interaction picture with respect to the detector , is given by fourier transform of the correlation function@xcite @xmath28 where @xmath29 is the detector dipole matrix element with positive - frequency part of the electric field is defined by @xmath30 and the negative - frequency part of the electric field is @xmath31^{\dag}$ ] . the quantity @xmath32 is the electric field per photon and @xmath33 is the effective volume of a cubic cavity resonator . we rewrite the correlation function , in eq . ( [ eqn : flou : spec:1 ] ) , in terms of the system operators using the methods described in refs . : @xmath34 here @xmath35 is the detector response function [ discussed in appendix [ sec : detector response ] ] . the average @xmath36 , in eq . ( [ eqn : flou : spec:11 ] ) , is given with respect to the stationary density matrix @xmath37 , where @xmath38 . using the cyclic property of trace , fluorescence spectrum can be written as @xmath39 here operator @xmath40 is defined as @xmath41 and @xmath42 . since @xmath43 and @xmath44 are operators in the system hilbert space and @xmath45=0 $ ] , the evolution of @xmath40 is determined by the hamiltonian of emitting system and radiation field , @xmath46 , in the absence of detector and can be computed without using the well - known qrt@xcite . we consider a driven two - level cavity - qed system with an excited state @xmath18 representing a single exciton , and a ground state @xmath47 with no exciton . the qd interacts with the cavity photons and a phonon reservoir which is coupled to the excited state @xmath18 , as shown in fig . [ fig : tls : rwa ] . the hamiltonian for the total system reads , @xmath48 where @xmath49 is the transition frequency of the two - level system and @xmath17 is the frequency of laser field . the photon ( phonon ) modes are represented by bosonic fields with frequencies @xmath13 ( @xmath50 ) with creation and annihilation operators @xmath11 ( @xmath51 ) and @xmath10 ( @xmath52 ) , respectively . the system - photon ( system - phonon ) coupling strength is given by @xmath12 ( @xmath53 ) , and @xmath54 ( rabi frequency ) is the coupling between the two - level system and laser field . the system operators are denoted by @xmath55 where @xmath56 and @xmath57 . the explicit time dependence in @xmath58 [ eq . ( [ eqn : full ptls h ] ) ] can be removed by going to a rotating frame and applying a rwa . we perform the rwa on both the driving term and the system - photon coupling term , in which we neglect the rapidly - oscillating terms and keeping only the time - independent part . the resulting rwa hamiltonian in the rotated frame is then @xmath59 where @xmath60 and @xmath61 are the detunings of atomic and cavity frequencies from the laser pump frequency @xmath17 . in general , operators in the rest frame are transformed to the rotating frame according to the following relations : @xmath62 due to presence of the intense laser field the two bare states are strongly coupled to each other and give rise to two dressed states . the dressed states can be written , in terms of the bare states , as : @xmath63 and @xmath64 with the mixing angle @xmath65 is given by @xmath66 , where @xmath67 is the dressed rabi frequency . .,scaledwidth=45.0% ] furthermore , following closely the discussing on ibm in ref . [ ] , we apply another transformation by using a canonical transformation @xmath68 , where @xmath69 is an anti - hermitian operator , to write the total hamiltonian in terms of the free and perturbed parts as , after transforming it to the dressed - state basis @xmath70 where the free part is @xmath71 and the perturbed part is given by @xmath72\label{eqn : trans : rad},\end{aligned}\ ] ] where @xmath55 , @xmath73 and @xmath74 . the polaron transformation introduces a frequency shift as : @xmath75 , where @xmath76 is the polaron shift . we have also performed a schrieffer - wolff transformation on the above hamiltonian to get rid of the energy exchange term ( @xmath77 lifetime process ) due to phonons , see appendix [ sec : schrieffer - wolff ] . the separation into pure dephasing and transition terms is determined by the form of system coupling , i.e. dephasing terms contain the diagonal coupling and transition terms contain the off - diagonal couplings . accordingly , fluorescence spectrum , in eq . ( [ eqn : flou : spec:12 ] ) , can be written in laplace domain using the dressed - state representation as a three - peak spectrum [ ref . ] : @xmath78_{s =- i\delta_{0 } } , \label{eqn : three - peak}\end{aligned}\ ] ] where @xmath79 is the probe detuning , we have used the relations @xmath80 , @xmath73 and @xmath81 . the polaron transformation does not affect the fluorescence spectrum in above expression as @xmath82 acts on the phonon mode hilbert space and commutes with the system operators . it can be seen that the first term in eq . ( [ eqn : three - peak ] ) gives the central peak whereas last two terms give rise to satellite peaks of the mollow - triplet , shown in fig . [ fig : dressed - states ] . we are interested in the influence of pure dephasing due to phonon interaction , which only affects the off - diagonal elements of the system operator . from here and what follows , we will study the effect of pure dephasing on the fluorescence spectrum , with a particular focus on one of the mollow - triplet sidebands ( stokes line ) . alternatively , when the width of each side peak @xmath83 , [ see eq . ( [ eqn : mar : decay : rate ] ) , below ] is small compared to the peak separation @xmath84 , we approximate the spectrum near the side peak centered at @xmath85 by , see fig . [ fig : dressed - states ] @xmath86 . \label{eqn : one - peak}\ ] ] in order to compute the spectrum , we will evaluate the dynamics of matrix element @xmath87 $ ] in the dressed - state representation with hamiltonian in the polaron frame , given by eq . ( [ eqn : polaron : hamiltonian ] ) . the radiation and phonon modes are decoupled from the system for times @xmath88 ( where @xmath89 is a time in the distant past ) , and prepared independently in the states described by density matrices @xmath90 , @xmath91 and @xmath92 , respectively . the interactions ( photons and phonons ) are switched on at this time @xmath93 , and the state of entire system is described by the full density matrix @xmath94 : @xmath95 where initial density matrix for photon is described by vacuum of the cavity modes @xmath96 and phonon modes are described by a canonical ensemble at temperature @xmath97 : @xmath98}. \label{eqn : phonon density matrix}\ ] ] we recall the operator @xmath40 , where @xmath99 is analogous to the density matrix operator with a modified initial condition , given by @xmath100 where stationary density matrix @xmath37 and hence @xmath101 account for conditions that accumulate between the system and interactions in the time interval @xmath102 $ ] . we choose an initial condition when exciton is in excited state @xmath18 given by @xmath103 , and evolves in the presence of pump laser . we switch on the detector at @xmath104 and subsequently calculate the dynamics of @xmath105 for @xmath106 , with initial condition given by the steady state density matrix accumulated between the time interval @xmath107 $ ] . we are interested in the dynamics of reduced system operator , after tracing over variables of photon and phonon modes ; @xmath108 . to study the dynamics of reduced system operator @xmath109 , we introduce a projection superoperator @xmath110 , defined by its action on an operator : @xmath111 . both operators , @xmath112 and @xmath40 , follow the same von - neumann equation , and can be written in form of exact nakajima - zwanzig gme@xcite , using @xmath113 @xmath114 where @xmath115 is the self - energy superoperator @xmath116 and @xmath117 is the full liouvillian superoperator , defined as @xmath118 $ ] and @xmath119 . we have used the properties of projection operator : @xmath120 and @xmath121 also introducing its complement @xmath122 . we can derive an equation of motion for @xmath40 analogous to the equation for @xmath112 [ eq . ( [ eqn : gme for density matrix ] ) ] . however , an additional term appears because @xmath123 , and we have assumed that the full density matrix operator is not separable for all times i.e. @xmath124 . the resulting motion equation for @xmath125 is then , @xmath126 where @xmath115 is defined in eq . ( [ eqn : self - energy superoperator ] ) and the last term in the above equation contains @xmath127 , i.e. the irrelevant part of @xmath101 is non - zero . when the radiation and phonon modes are described by eqs . ( [ eqn : photon density matrix ] ) and ( [ eqn : phonon density matrix ] ) , the projection operator @xmath110 follows some useful identities @xmath128 we apply above identities ( [ eqn : iden1])-([eqn : iden4 ] ) , and perform the partial traces on eqs . ( [ eqn : gme for density matrix ] ) and ( [ eqn : gme for pseudo density matrix ] ) over variables of radiation and phonon modes to obtain an equation of reduced system operators , @xmath129,\label{eqn : reduced : self : energy}\\ \phi_{s}(t)&=-i\mathrm{tr}_{r}\mathrm{tr}_{p}\bigg[l_{v}e^{-iqlt}q\omega(0)\bigg],\label{eqn : irr : part}\end{aligned}\ ] ] where @xmath130 and @xmath131 are reduced self - energy and irrelevant part superoperators , respectively , and @xmath132 is the reduced system operator . comparing eqs . ( [ eqn : eom : rhos ] ) and ( [ eqn : eom : omegas ] ) , we found that the first two terms are identical but there is an additional irrelevant part , @xmath131 , is present in the equation for @xmath109 . this term vanishes in a markov approximation and usual qrt can be applied to find the system correlation@xcite . in addition to this , we have also assumed that the full density matrix is _ not _ separable for all times which is another reason that qrt is no longer valid in the present case . furthermore , for a non - markovian equation the irrelevant term is non - zero and usual qrt can not be used to compute the system correlation in this case@xcite . the equation for off - diagonal matrix element , @xmath133 , is coupled to both diagonal @xmath134 and off - diagonal @xmath135 elements of the reduced system operator , and can be written in the form @xmath136 with a non - zero irrelevant part matrix element expressed as ( see appendix [ sec : irrelevant - part ] ) @xmath137_{+-,+-}=\bigg[-tr_{r}tr_{p}l_{v}\,e^{-iqlt}\bigg(\frac{1}{0^{+}+iql}l_{v}\bigg)\rho_{r}(t_0)\rho_{p}(t_0)\bigg]_{+-,+-}.\ ] ] the self - energy superoperator can be decomposed into three terms as @xmath138 where the first and second terms in above expression are pure dephasing ( @xmath139 pure dephasing time ) processes due to photon and phonon couplings , respectively , and the third term gives rise to transition ( @xmath77 lifetime ) due to radiative modes coupling . we treat the dot - cavity coupling self - energy within born - markov approximation to second - order in perturbation liouvillain @xmath140 , see appendix [ sec : photon - self - energy ] . applying the continuum of modes for cavity given by lorentzian density of state [ eq . ( [ eqn : dos : cavity ] ) ] , we find the self - energy matrix elements in their laplace domian , defined as @xmath141 , @xmath142 @xmath143 where @xmath144 is the detuning of cavity from laser pump frequency and @xmath145 is cavity bandwidth . here , born approximation is justified by finding that the higher order terms in reduced self - energy are suppressed by a small parameter @xmath146 . similarly for phonon modes , applying a continuum of modes [ see appendix [ sec : phonon - self - energy ] ] for the deformation potential coupling mechanism@xcite @xmath147 , where @xmath148 is the phonon coupling parameter in the units of @xmath149 and @xmath150 is the phonon cut - off frequency , we obtain @xmath151 here @xmath152 is bose function . in above expression , the dot - phonon interaction self - energy is evaluated using exact approach within non - markovian limit and it is found that born approximation is exact in this case , see appendix [ sec : exact : self : energy ] . furthermore , the reduced self - energy matrix elements couple to the populations can as well be written within usual born approximation @xmath153 @xmath154 and similarly for the coherence @xmath135 , @xmath155 equation for the coherence in eq . ( [ eqn : full : eom ] ) contains both diagonal and off - diagonal elements of the self - energy superoperator , some of them are fast moving compare to others . in next section , we will perform a secular approximation@xcite to get rid of the fast oscillating terms . the secular approximation consists in neglecting the fast oscillating terms in markov equation - of - motion , and the equation for @xmath133 [ eq . ( [ eqn : full : eom ] ) ] is decoupled from populations and coherence within a secular approximation@xcite . here , we consider a general equation of motion for the operator @xmath40 , without irrelevant part matrix elements : @xmath156 and would like to perform the secular approximation in order to get rid of the fast oscillating terms . to this end , introducing a rotating frame @xmath157 where @xmath158 is the total frequency shift given implicitly by the expression @xmath159 in the weak coupling regime , @xmath160 , the frequency shift to the leading order in @xmath158 @xmath161 . \label{eqn : total : freq : shift1}\ ] ] here , the purpose of introducing a rotating frame is to get rid of all oscillating parts from @xmath162 and hence obtain an equation for @xmath163 @xmath164 where @xmath165 . due to presence of the oscillatory exponential in the last two terms in rhs of the above equation , we decompose it into slow - varying and fast - oscillating parts as : @xmath166 where s and f stand for slow and fast , respectively . carrying out markov approximation on slow term by replacing : @xmath167 , @xmath168 , and finally extending the upper limit of integration to infinity , we obtain @xmath169 where @xmath170 condition for validity of markov approximation : @xmath171 decays on a time scale @xmath172 which is the decay time of @xmath173 and given by the relation@xcite @xmath174 on substituting eqs . ( [ eqn : dp : pm ] ) and ( [ eqn : trans : pm ] ) in the above inequality and for @xmath175 and @xmath176 , it leads to the condition @xmath177 and similarly for the phonon coupling @xmath178 . similarly , for the fast term @xmath179 and the condition for validity of markov approximation : @xmath180 which leads to a similar condition @xmath177 . for large @xmath181 and due to presence of the highly oscillatory exponential @xmath182 , the effects of terms @xmath183 will eventually average out to the smaller values compared to @xmath184 . therefore , in secular approximation when @xmath185 we neglect the fast oscillating terms in markovian approximation since this term will oscillate fast and average out to a smaller value compared to the slow term . on substituting for @xmath186 from eq . ( [ eqn : trans : mp ] ) and for @xmath187 , @xmath160 , we get an explicit condition for the validity of secular approximation as : @xmath188 . in the similar manner , we can also neglect the contribution form @xmath189 and @xmath190 from eq . ( [ eqn : full : eom ] ) . going back to lab frame and within secular approximation , we obtain the expression in laplace transform : @xmath191 with initial condition expressed in terms of reduced self - energy matrix elements given by eqs . ( [ eqn : trans : pp ] ) and ( [ eqn : trans : mm ] ) , see appendix [ sec : initial - condition ] : @xmath192 the irrelevant part matrix elements , due to photon @xmath193 and phonon @xmath194 , both are identically zero under a markov approximation . on substituting for @xmath195 from eq . ( [ eqn : fully : mar : eqn ] ) in the expression ( [ eqn : one - peak ] ) , we obtain an expression for one - peak markovian spectrum @xmath196 which is a lorentzian line centered at @xmath197 with a width given by @xmath198 , frequency - shift @xmath158 and decay @xmath198 are given by eqs . ( [ eqn : total : freq : shift1 ] ) and ( [ eqn : mar : decay : rate ] ) , respectively , and the pre - factor is given by @xmath199 here , we have substituted the expression for the self - energy matrix elements given by eqs . ( [ eqn : trans : pp ] ) and ( [ eqn : trans : mm ] ) . expression for the lineshape in eq . ( [ eqn : fully : mar : eqn ] ) is markovian with respect to both photon and phonon interactions . however , we are interested in the non - markovian regime with respect to phonon coupling and an equation for @xmath105 always contains an extra small term @xmath200 due to non - markovian interaction , known as irrelevant part matrix element . assuming that the irrelevant part is associated with a smallness in the present problem and in order to get the further insight , we will estimate the typical size of its contribution and find a regime where non - markovian correction is dominant compared to its irrelevant part contribution . equation for @xmath105 in laplace domain with its irrelevant part matrix element can be written as , @xmath201\omega_{+-}(0 ) . \label{eqn : laplace : irrelevant : part}\ ] ] the smallness of irrelevant part compared to non - markovian self - energy due to phonon interaction can be justified by the inequality given by @xmath202 and we only keep the contribution from self - energy matrix element . furthermore , expanding eq . ( [ eqn : laplace : irrelevant : part ] ) in the powers of self - energy and ignoring the higher order terms , we have @xmath203\omega_{+-}(0),\ ] ] and assuming that irrelevant part gives rise to a small contribution and comparing it with self - energy contribution leads to the following inequality , @xmath204 the one - peak spectrum is centered around @xmath205 , with a width mostly dominated by markovian decay rate @xmath198 , estimating the size of above inequality around @xmath206 , and it is found that irrelevant part matrix element is always suppressed by a small parameter , @xmath207 , compared to self - energy matrix element contribution , also see appendix [ sec : irrelevant - part ] . following the above discussion , we neglect the irrelevant part from the equation for @xmath105 and after going back to lab frame we have an equation written in laplace domain , @xmath208+\gamma_{r}+\gamma_{p}(\delta_{0})}. \label{eqn : non : mar : phonon}\ ] ] above expression is markovian with respect to photon coupling but non - markovian in terms of phonon interaction , where markovian frequency shift ( @xmath209 ) and decay rate ( @xmath210 ) are given by @xmath211_{s =- i\omega_{r}'},\\ \gamma_{r}&=-{\rm im}[\sigma_{+-,+-}^{dr}(s)+\sigma_{+-,+-}^{sr}(s)]_{s =- i(\omega_{r}'+\delta\omega)}.\end{aligned}\ ] ] similarly , non - markovian frequency - dependent shift ( @xmath212 ) is expressed as @xmath213_{s =- i\delta_{0}}\ ] ] and dephasing ( @xmath214 ) is given by @xmath215_{s =- i\delta_{0}}.\ ] ] on substituting for @xmath195 from eq . ( [ eqn : non : mar : phonon ] ) in the expression ( [ eqn : one - peak ] ) , we obtain an expression for the one - peak non - markovian spectrum @xmath216}{[\delta_{0}-\omega_{r}'-\delta\omega_{r}-\delta\omega_{p}(\delta_{0})]^{2}+[\gamma_{r}+\gamma_{p}(\delta_{0})]^{2}},\ ] ] where pre - factor @xmath217 is given by eq . ( [ eqn : prefactor ] ) . it should be noted here that we have not performed born - markov approximation in terms of phonon interaction . in markovian regime frequency shift @xmath212 and dephasing @xmath214 are replaced by their @xmath197 frequency parts and give rise to an exponential decay and hence to a lorentzian line centered at @xmath197 with a width given by @xmath198 . whereas in the non - markovian regime frequency - shift and dephasing due to phonon interaction are frequency dependent and lead to a non - markovian ( non - exponential ) decay giving rise to non - lorentzian features in the lineshape . we apply above obtained theoretical results to inas / gaas qds , and use the cavity and phonon parameters given in refs . . in this section , we plot and analyze the results obtained in the previous sections for different parameter regimes . typical phonon parameters for gaas are obtained from refs . : phonon cut - off @xmath218 and coupling @xmath219 . for the laser and cavity , we choose following parameters@xcite @xmath220 , @xmath221 , @xmath222 . we have varied the other parameters in the plots and explained along with the figures . in fig . [ fig : self ] , we plot the frequency - shift and dephasing as a function of probe - detuning for different values of phonon bath temperatures . we observe that the effect is strongly temperature - dependent . it increases linearly in the high temperature limit and vanishes for small temperatures . in the present parameter regime for a typical gaas qd , the dominant contribution from phonon interaction is mainly due to the frequency - shift , see fig . [ fig : self ] . here , we have assumed that the dot and cavity are resonant and set @xmath223 . [ cols="^,^ " , ] we also observe that the lamb shift and additional broadening , due to phonon interaction , vanish for the case of resonant pump laser . this can be explained in the _ dressed - state _ representation , where both energy - levels of interest , correspond to one of the transitions ( stokes line ) , are coupled asymmetrically to the phonons . for the zero detuning they become degenerate leading to the vanishing frequency - shift and dephasing . we have also observed a little shifts in the peak positions due to strong dependence of frequency - shift on the probe - detuning . we have discussed the dynamics of a driven cavity - qed system coupled to a 3d acoustic phonon reservoir with non - markovian pure dephasing mechanism . we have observed highly modified non - lorentzian features in the associated spectrum due to phonon interaction , and for solid - state systems it can only be described within non - markovian regime . we have shown that quantum regression theorem is not valid in this case because of a non - zero irrelevant part , and also the full density matrix is _ not _ separable for all times . the system correlation is solved without using the quantum regression theorem beyond usual born - markov approximation . we have obtained the analytical formulas for both markovian and non - markovian one - peak lineshapes in terms of the model parameters . both expressions look like lorentzian lines centered around @xmath224 , but this analogy is not correct as shift and dephasing are also probe - dependent in case of non - markovian lineshape . we have investigated the one - peak spectrum of the mollow - triplet for different values of phonon bath temperature for the resonant dot and cavity . we have found vanishing shift and dephasing when laser is resonant with the dot , because levels are equally and asymmetrically coupled to phonon modes rendering them degenerate for a resonant pump laser . we have derived an exact form of najakjima - zwanzig generalized master equation and showed that markov and non - markov solutions give significantly different results . we show that non - markovian contribution is significantly large and can be clearly seen in the spectrum . a distinct narrowing of the sideband has been reported which is contrary to the recent results leading to broadening in the presence of phonons . we have shown that the frequency - dependent shift has strong temperature dependence which causes different features like narrowing and asymmetry in the lineshape . this procedure can also be used to systematically account for features in optical spectra of a general multi - level system due to genuine non - markovian dynamics . ak acknowledge financial support from the icelandic research fund rannis , cifar , canada , the national key research and development program of china ( grant no . 2016yfa0301200 ) , and the nfsc grants ( no 11574025 and no . u1530401 ) . ak thanks bill coish , sigurdur i. erlingsson , stefano chesi , li - jing jin , and tilen cadez for the useful discussions and feedback . the detector response function in eq . ( [ eqn : flou : spec:11 ] ) @xmath225 where @xmath226 . applying the continuum of modes after replacing the sum via integral @xmath227 and using the well - known formula @xmath228 ( @xmath229 indicates the principal part ) one can write the square of detector response function in terms of principal part and delta function as @xmath230\nonumber\\ & \times\int_{0}^{\infty}d\nu\,d(\nu)\|g^{d}(\nu)\|^{2}\bigg[i\mathcal{p}\left(\frac{1}{\nu}\right)+\pi\delta(\nu)\bigg],\end{aligned}\ ] ] where @xmath231 and @xmath232 are photonic density of states of cavity and open space , respectively . the photonic density of states of the cavity is described by a lorentzian density of states given by eq . ( [ eqn : dos : cavity ] ) . hamiltonian @xmath233 resulting from the polaron transformation , is @xmath234\nonumber\\ & -{\bf c.s}\,\sum_{q}\lambda_{q}(\sigma_{+-}+\sigma_{-+})(b_{q}+b_{q}^{\dagger})\nonumber\\ & -\sum_{q}\frac{\lambda_{q}^{2}}{4\omega_{q}}+{\bf c.s}\,\sum_{q}\frac{\lambda_{q } ^{2}}{\omega_{q}}(\sigma_{+-}+\sigma_{-+}).\end{aligned}\ ] ] last two terms in above hamiltonian commute with rest of the hamiltonian and can be igonred within a secular approximation for large @xmath235 . we get rid of the energy exchange process due to phonon interaction using leading - order schrieffer - wolff transformation@xcite and start from the full hamiltonian : @xmath236 where individual terms are defined as @xmath237,\\ v_{2}&=-{\bf c.s}\,\sum_{q}\lambda_{q}(\sigma_{+-}+\sigma_{-+})(b_{q}+b_{q}^{\dagger}).\end{aligned}\ ] ] we apply a transformation , @xmath238 , generated by an anti - hermitian operator @xmath239 to eliminate the transition terms to the first order . using baker - campbell - hausdorff formula and expanding @xmath240 in the powers of @xmath241 , we obtain @xmath242+[a , v_{2}]+\frac12[a,[a , h]]+ ... \ ] ] in order to get rid of transition term @xmath243 to leading order , we set @xmath244 $ ] , where @xmath241 can be written as @xmath245 and @xmath246 $ ] . here @xmath241 is of order of transition term @xmath243 . substituting for @xmath241 , we obtain hamiltonian up to the second or higher order in @xmath243 @xmath247+ ... \ ] ] using the definitions of @xmath248 and @xmath243 , we obtain the expression for a as : @xmath249 therefore , the transformed hamiltonian to the first order in the transition terms due to phonon can be well approximated and written as free and perturbed parts as @xmath250 where the free part is @xmath71 and perturbed parts is given by @xmath72,\end{aligned}\ ] ] which are eqs . ( [ eqn : polaron : hamiltonian])-([eqn : trans : rad ] ) in the main text . the irrelevant part is given by eqn . ( [ eqn : irr : part ] ) can be written within born approximation after transforming to laplace domain @xmath251 in particular , we want the matrix element , @xmath252 due to phonon interaction , which can be simplified and written as @xmath253_{+-},\ ] ] where irrelevant part of the stationary density matrix can be found@xcite using gme discussed in sec . [ sec : gme ] @xmath254_{+-}=-i\lim_{s\rightarrow 0}\frac{s}{s+i(\omega_{r}'+l_{p})}l_{y}^{+}\rho_{p}(t_0)\,[\rho_{s}(s)\sigma_{ab}]_{+-}. \label{eqn : irrelevant part matrix element}\ ] ] the propagators in eqn . ( [ eqn : irrelevant part matrix element ] ) has no poles at @xmath255 , where @xmath256 is a positive infinitesimal . after performing the limit in the above expression , we substitute for @xmath257_{+-}$ ] in the expression for @xmath252 , to obtain and expression for the irrelevant term as @xmath258 where @xmath259_{+-}$ ] , see eqs . ( [ eqn : initial : condition : matrix ] ) and ( [ eqn : initial : condition : matrix : element ] ) . above expression can be written in terms of irrelevant part matrix element in the problem as @xmath260 solving above expressions and applying the continuum of modes , we found that irrelevant part matrix element will be suppressed by @xmath261 compared to the self - energy matrix element due to phonon interaction . in this limit , contribution from irrelevant part can be neglected compared to the contribution from the non - markovian self - energy . the reduced self - energy superoperator @xmath130 in eq . ( [ eqn : reduced : self : energy ] ) can be transformed in to laplace domain and using @xmath262 , we obtain @xmath263 dropping @xmath264 from the exponential in eq . ( [ eqn : reduced : self : energy ] ) will not affect the final expression @xcite . cross terms in above expression will vanish because @xmath265 and @xmath266 act on an operator in the radiation mode hilbert space whereas @xmath267 act on phonon hilbert space and will not contribute in the final trace . moreover , the cross terms between @xmath265 and @xmath266 will give rise to off - block diagonal matrix elements in self - energy matrix and will be neglected within secular approximation , see sec . [ sec : secular : approximation ] in main text . the self - energy superoperator in eq . ( [ eqn : decomposed : self : energy ] ) can be decomposed into three different parts as : @xmath268 where first and second terms give rise to pure dephasing ( @xmath139 process ) due to radiation and phonon modes , respectively , whereas last term in above expression leads to transition ( @xmath77 process ) due to radiation modes coupling . free propagator in reduced self - energy expression can be expanded in the powers of interacting liouvillian @xmath140 as @xcite @xmath269 because of the form of couplings in present model only even powers @xmath270 in above expression will survive in the final trace . in order to find the matrix elements of the self - energy superoperator , we write the superoperators in matrix form in dressed - state basis as @xmath271= \left ( \begin{array}{cccc } 0 & 0 & 0 & 0\\ 0 & 0 & 0 & 0\\ 0 & 0 & \omega_{r } ' & 0\\ 0 & 0 & 0 & -\omega_{r } ' \end{array } \right),\ ] ] where @xmath272_{\alpha\beta,\gamma\delta}=tr\lbrace\ket{\beta}\bra{\alpha}s\ket{\gamma}\bra{\delta}\rbrace$ ] and @xmath273 . in the dressed state basis , non - interacting liouvillian is diagonal and can be inverted to write as @xmath274 blocks @xmath275= \left ( \begin{array}{cc } g_{\parallel}(s ) & 0\\ 0 & g_{\perp}(s ) \end{array } \right),\ ] ] where parallel block is @xmath276= \left ( \begin{array}{cc } \frac{1}{s+i(l_{r}+l_{p } ) } & 0\\ 0 & \frac{1}{s+i(l_{r}+l_{p } ) } \end{array } \right),\ ] ] and perpendicular block is given by @xmath277= \left ( \begin{array}{cc } \frac{1}{s+i(\omega_{r}'+l_{r}+l_{p } ) } & 0\\ 0 & \frac{1}{s+i(-\omega_{r}'+l_{r}+l_{p } ) } \end{array } \right).\ ] ] in the similar fashion , we can find other matrices as well @xmath278= \left ( \begin{array}{cccc } l_{x(y)}^{- } & 0 & 0 & 0\\ 0 & l_{x(y)}^{- } & 0 & 0\\ 0 & 0 & l_{x(y)}^{+ } & 0\\ 0 & 0 & 0 & -l_{x(y)}^{+}\\ \end{array } \right),\ ] ] here we have defined new liouvillian for commutation and anti - commutation relations : @xmath279_{\pm}$ ] , where operators @xmath280 and @xmath281 are given as @xmath282 furthermore , we also find the matrix for superoperator @xmath266 in the dressed - state basis @xmath283= \left ( \begin{array}{cccc } 0 & 0 & -z_{r}^{\dag } & z_{l}\\ 0 & 0 & z_{l}^{\dag } & -z_{r}\\ -z_{r } & z_{l } & 0 & 0\\ z_{l}^{\dag } & -z_{r}^{\dag } & 0 & 0 \end{array } \right ) , \ ] ] where we have defined the operators for left and right multiplications as : @xmath284 and @xmath285 is an operator in the radiation mode hilbert space . reduced self - energy matrix elements of interest can be calculated according to eq . ( [ eqn : full : eom ] ) in the main text . we apply a continuum of modes for the cavity density of states given by a lorentzian spectrum @xcite , @xmath286 where @xmath144 is the detuning of cavity from laser pump frequency , and @xmath145 is cavity bandwidth . reduced self - energy matrix elements due to radiation mode coupling giving rise to transition are given as @xmath287 and self - energy matrix element responsible for pure dephasing due to cavity coupling can be found as @xmath288 for a large band - width cavity , coupling to its radiative modes to system is treated under markov approximation and above self - energies are replaced by their @xmath289 frequency parts , refer main text for details . similarly , we apply a continuum of modes for 3-d acoustic phonons@xcite with an exponential cut - off at @xmath290 @xmath291 we obtain the expression for self - energy in laplace transform as @xmath292 where @xmath293 is the phonon coupling paramter in the units of @xmath294 and @xmath152 is bose function . on further simplification , above self - energy matrix element can be decomposed into real and imaginary parts after setting @xmath295 , where @xmath296 is detuning of the probe from the pump laser frequency , as : @xmath297 where @xmath298 $ ] is the frequency - shift and @xmath299 $ ] is the dephasing due to phonon interaction . in the previous section , we have computed the self - energy matrix element for phonon interaction to the second order in born approximation . in this section , we will discuss an equation - of - motion method to find the phonon interaction self - energy for all orders in perturbed liouvillain due to phonons @xmath300 beyond born approximation , and show that exact approach recovers the result obtained within born approximation . using the general form of superoperators matrices , the expression for self - energy matrix element due to phonons can be written in laplace domain as : @xmath301 also in the time domain , we have @xmath302}_{\mathcal{c}(t)}. \label{sigma_phonon_exact}\ ] ] on further simplification , one obtains @xmath303_{+}\rho_{p}(0)\biggr]\nonumber\\ & = 2\langle x_{p}x_{p}(t)\rangle+2\langle x_{p}(t)x_{p}\rangle,\end{aligned}\ ] ] where @xmath304 above expression gives rise to a differential equation @xmath305 which can be written as @xmath306 introducing : @xmath307 , we obtain an equation of motion for @xmath308 @xmath309_{+ } \label{eqn : eom in ques}\ ] ] where @xmath310 solving for one @xmath311 , the solution for @xmath308 takes the form @xmath312 where @xmath313 using eqns . ( [ eqn : eom with hq ] ) and ( [ eqn : eom uq ] ) , we have ( for one @xmath311 ) @xmath314 where @xmath315 introducing @xmath316 and taking the time derivative , one can find an expression @xmath317\tilde{u}_{q}(t).\ ] ] using the shifting operator @xmath318 above expression can be written as @xmath319 this implies @xmath320 multiplying by @xmath321 on both sides @xmath322 solving for @xmath323 @xmath324 using @xmath325 and @xmath326 , we obtain @xmath327 similarly , for @xmath328 @xmath329 substituting for @xmath330 and @xmath328 , also using @xmath331 , we obtain an expression for @xmath332 @xmath333 recalling eq . ( [ eqn : eom in ques ] ) : @xmath334 and substituting for @xmath332 , we have @xmath335.\end{aligned}\ ] ] above expression can be solved for @xmath336 , because other modes do not contribution in the final trace and do not conserve the particle number for the different modes . on further simplification and doing some algebraic manipulation , one can obtain a simplified expression as @xmath337\ ] ] and similarly for @xmath338 @xmath339.\ ] ] considering the factor in parenthesis which is common in both expressions , we have @xmath340 where we have defined : @xmath341 using baker - campbell - hausdorff formula : @xmath342 similarly for @xmath343 , @xmath344 substituting for @xmath343 and @xmath345 in the expression for @xmath346 , and then substituting for @xmath346 in the expressions for @xmath347 and @xmath338 , we perform the final trace to obtain the following expressions @xmath348 after substituting the expressions for @xmath347 and @xmath338 in the expression for reduced self - energy matrix element and performing the final trace , we obtain @xmath349 applying the continuum of modes and going back to laplace domain , @xmath350 above expression recovers the result obtained for self - energy matrix element within born approximation , which is eq . ( [ eqn : phonon : self : energy ] ) in the main text . in this section , we will discuss the stationary density matrix and find the initial condition for the operator @xmath40 given in eq . ( [ eqn : corr : initial : condition ] ) . the stationary density matrix @xmath37 accounts for conditions that accumulate between the system and interactions in the time interval @xmath102 $ ] . in this time interval , @xmath107 $ ] , the stationary density matrix is given by its value at @xmath104 and we can replace @xmath351 in eq . ( [ eqn : eom : rhos ] ) , to obtain @xmath352 changing integration variable : @xmath353 to obtain @xmath354 one can extend the upper limit of above integration to infinity by setting @xmath355 , and solving the differential equation to obtain @xmath356t_0}\rho_{s}(t_0),\ ] ] and taking laplace transform on both sides , we obtain @xmath357 here laplace transform is defined as @xmath141 . we choose an initial condition when exciton is in excited state @xmath18 given by @xmath103 , and evolves in the presence of pump laser . after performing a secular approximation and using the definition of stationary limit @xmath358 one can find all the elements of stationary density matrix operator as : @xmath359 where @xmath360 . similarly , @xmath361 recalling , the initial condition for operator @xmath40 given by eq . ( [ eqn : corr : initial : condition ] ) : @xmath362 and matrix element of interest can be extracted after performing the trace over phonon and photon modes , and substituting for stationary density matrix element @xmath363 , @xmath364 dynamics of the operator @xmath40 is evaluated using hamiltonian given by eq . ( [ eqn : polaron : hamiltonian ] ) , with an initial condition given by above expression which is eq . ( [ eqn : initial : condition ] ) in the main text . 43ifxundefined [ 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 _ _ ( ) `` , '' ( , , ) chap . , pp . link:\doibase 10.1103/physreva.51.2738 [ * * , ( ) ] link:\doibase 10.1103/physreva.57.120 [ * * , ( ) ] link:\doibase 10.1038/nature00912 [ * * ( ) , 10.1038/nature00912 ] link:\doibase 10.1126/science.1173684 [ * * , ( ) ] link:\doibase 10.1103/physrevb.93.241306 [ * * , ( ) ] link:\doibase 10.1038/ncomms9204 [ * * ( ) , 10.1038/ncomms9204 ] link:\doibase 10.1364/oe.21.004382 [ * * , ( ) ] link:\doibase 10.1103/physreva.73.053807 [ * * , ( ) ] http://stacks.iop.org/1367-2630/12/i=11/a=113042 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.247402 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.115309 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.217401 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.241304 [ * * , ( ) ] link:\doibase 10.1103/physrevb.65.195313 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.075309 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.118.063601 [ * * , ( ) ] link:\doibase 10.1103/physreva.95.013834 [ * * , ( ) ] link:\doibase 10.1103/physrevb.70.195340 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.104.157401 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevb.95.075412 [ * * , ( ) ] link:\doibase 10.1103/physrevx.6.031004 [ * * , ( ) ] link:\doibase 10.1103/physreva.93.022119 [ * * , ( ) ] @noop _ _ ( , , ) link:\doibase 10.1103/physrevlett.99.187402 [ * * , ( ) ] @noop _ _ ( , , ) link:\doibase 10.1103/physrev.188.1969 [ * * , ( ) ] @noop _ _ ( , , ) link:\doibase 10.1103/physrevb.95.035302 [ * * , ( ) ] link:\doibase 10.1103/physrevb.94.161302 [ * * , ( ) ] @noop _ _ ( , , ) link:\doibase 10.1103/physreva.77.043836 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/s0030-4018(00)00622-2 [ * * , ( ) ] http://stacks.iop.org/1367-2630/18/i=8/a=083038 [ * * , ( ) ] @noop _ _ ( , , ) @noop _ _ ( , , ) link:\doibase 10.1103/physrev.149.491 [ * * , ( ) ] link:\doibase 10.1103/physrevb.71.035318 [ * * , ( ) ] link:\doibase 10.1103/physreva.85.062323 [ * * , ( ) ]	we present a detailed theoretical study of non - markovian dynamics in the fluorescence spectrum of a driven semiconductor quantum dot ( qd ) , embedded in a cavity and coupled to a three - dimensional ( 3d ) acoustic phonon reservoir . in particular , we investigate the effect of pure dephasing on one of the side - peaks of the mollow - triplet spectrum , expressed in terms of the off - diagonal element of the reduced system operator . the qd is modeled as a two - level system with an excited state representing a single exciton , and ground state represents the absence of an exciton . coupling to the radiative modes of the cavity is treated within usual born - markov approximation , whereas dot - phonon coupling is discussed within non - markovian regime beyond born approximation . using an equation - of - motion technique , the dot - phonon coupling is solved exactly and found that the exact result coincides with that of obtained within born approximation . furthermore , a markov approximation is carried out with respect to the phonon interaction and compared with the non - markovian lineshape for different values of the phonon bath temperature . we have found that coupling to the phonons vanishes for a resonant pump laser . for a non - resonant pump , we have characterzied the effect of dot - laser detuning and temperature of the phonon bath on the lineshape . the sideband undergoes a distinct narrowing and aquires an asymmetric shape with increasing phonon bath temperature . we have explained this behavior using a dressed - state picture of the qd levels .
big bang nucleosynthesis ( bbn ) theory predicts the abundances of the light elements , deuterium ( d ) , helium ( 3 and 4 ) , and lithium ( 7 ) produced during the first three minutes after the big bang . anisotropies detected in the cosmic microwave background ( cmb ) contain information about the universe at the time of last scattering , about 300,000 years after the big bang . both of these pillars of cosmology , probing two different epochs in the early universe offer independent determinations of the cosmic baryon density . comparing their results will provide a fundamental test of cosmology . there is tentative agreement between current obervations of the cmb anisotropies and the light element abundances , and their respective baryon density determinations adding confidence to the cosmic `` standard model '' @xcite . figure 1 . bbn predictions table 1 . most important reactions and their uncertainties . [ cols="<,>",options="header " , ] in standard bbn , primordial abundances are sensitive to only one parameter , the baryon - to - photon ratio , @xmath0 . light element observations are used to determine their primordial abundances @xcite . these in turn , are convolved with bbn theory to determine the baryon density . the uncertainties in the theory lie entirely with the nuclear reaction cross sections that govern the formation of the light nuclei @xcite . figure 1 shows the 95% confidence predictions of bbn theory . combined with observational constraints , shown with dark boxes , they pick out an @xmath1 . current cmb measurements by dasi and boomerang @xcite tentatively agree with this baryon density , as shown by the vertical band . with the advent of precision measurements of the anisotropies in the cmb by map and planck , we are obliged to go out of our way to improve both theory and observations in order to make the best comparison between bbn and the cmb . the reactions listed in table 1 dominate the uncertainties in the light element abundance predictions from bbn . the uncertainties listed are from @xcite , see also @xcite for different and complimentary methods . it is important to measure these key reactions to less than 4% so that bbn theory s predictions are more precise , making bbn a sharper probe of nuclear and particle astrophysics @xcite . besides improving the nuclear reaction uncertainties , it is important to reduce the errors in the light element abundance observations . a large number of 4 and 7 observations exist from extragalactic hii regions and the atmospheres of population ii stars in the galactic halo , respectively . both of these abundance determinations suffer from large systematic uncertainties . improvements in theoretical models of these regions are needed to reduce these systematics , and possible new obervational stageties should be attempted . deuterium however , consists of only a handful of observations . this small sample prevents us from analysing systematic uncertainties and exploring possible trends . only with more observations in high - redshift damped lyman-@xmath2 systems will d be put on the same statistical footing as 4 and 7 . with the direct comparision of the baryon densities derived from bbn and the cmb , we will fundamentally test the very framework cosmology is built upon . assuming concordance is established we can use bbn and the cmb together to place stronger constraints on physics @xcite . with the cmb baryon density , we can predict the light element abundances with high precision . using the observations of light element abundances we can address their evolution in universe . we can test new particle physics ( reviewed in @xcite ) and non - standard cosmology ( e.g. dark radiation @xcite ) . with new analyses of observational data determining the primordial light element abundances and updated nuclear cross section data , we will be able to perform precision cosmology . i would like thank timothy beers , sean ryan , jason prochaska , john norris , and scott burles for useful discussions about the light element observations . fields and k.a . olive , apj 506 ( 1998 ) 177 [ he ] . olive and e. skillman , new astron . 6 ( 2001 ) 119 . izotov and t.x . thuan , apj 500 ( 1998 ) 188 . + a. peimbert , m. peimbert and v. luridiana , apj 565 ( 2002 ) 668 . j.m . omeara _ et al_. , apj 552 ( 2000 ) 718 [ d ] . + d. tytler _ et al_. , phys . 333 ( 2000 ) 409 . + m. pettini and d.v . bowen , apj 560 ( 2001 ) 41 . + s. dodorico , m. dessauges - zavadsky and p. molaro , a&a 368 ( 2001 ) l21 .	big bang nucleosynthesis ( bbn ) and the cosmic microwave background ( cmb ) anisotro- pies independently predict the universal baryon density . comparing their predictions will provide a fundamental test on cosmology . using bbn and the cmb together , we will be able to constrain particle physics , and predict the primordial , light element abundances . these future analyses hinge on new experimental and observational data . new experimental data on nuclear cross sections will help reduce theoretical uncertainties in bbn s predictions . new observations of light element abundances will further sharpen bbn s probe of the baryon density . observations from the map and planck satellites will measure the fluctuations in the cmb to unprecedented accuracy , allowing the precise determination of the baryon density . when combined , this data will present us with the opportunity to perform precision cosmology .
the kondo effect is a paradigm of strong correlation in condensed matter physics.@xcite this effect was first discovered in metals with magnetic impurity where the resistivity enhances with decreasing temperature below a characteristic temperature known as kondo temperature @xmath1 . such anomaly behavior was theoretically explained by j. kondo , hence the name kondo effect " , as the exchange interaction of the itinerant electrons with a localized spin state.@xcite soon , it was realized that the same correlation can dominate the low temperature transport properties of quantum dots with strong coulomb interaction.@xcite contrary to metals , the kondo effect leads to the enhancement of conductivity of the quantum dots . the advance of nanotechnology has experimentally demonstrated the kondo effect in artificial impurity systems ( semiconductor quantum dots@xcite , carbon nanotube@xcite or molecular conductors@xcite ) . the observation of kondo effect in these artificial quantum impurities has provided us the opportunity to study correlation phenomena by tuning the relevant parameters or driving the system out of equilibrium by either static or alternate magnetic or electric field . however , comparing with the great achievement of experiments , the theoretical understanding of kondo effect , especially at nonequilibrium , is far from adequate . this difficulty is due to the interplay of strong coulomb interaction and nonequilibrium which makes the exact solution of the kondo problem in nonequilibrium impossible . for the theoretical studies of kondo effect , several techniques , such as equation of motion of green s function@xcite , bosonization technique@xcite , and numerical renormalization group method@xcite , with their own advantages and limitations are developed . up to now , a large literature has been devoted to study the transport properties of kondo systems . most of these studies focused on the spectral density , linear or differential conductance and the current . nowadays , there is increasing interest in the study of the fluctuation of current , i.e. the current noise@xcite , in the kondo regime . the reasons for these studies are roughly two folds . on one hand , current noise measurement is expected to expose information about the quantum nature of electron transport in kondo regime . for kondo dot , the kondo correlation is quite sensitive to the external electric or magnetic field . it is thus difficult to get some key information such as the kondo temperature or the spectral density by direct experimental measurement . new setups or novel characteristic tools are needed to get the desired information . the current noise of kondo dot has shown its potential to provide information which is beyond the reach of traditional transport measurement . for instance , meir and golub studied the shot noise of kondo dot by several complementary approaches@xcite . they proposed to estimate @xmath1 by measuring the noise behavior of kondo dot . in addition , sindel _ et al . _ pointed out that the current noise can give an estimate of the kondo resonance in the equilibrium spectral density.@xcite on the other hand , as a signature of many - particle correlation , the current noise in kondo regime behaves quite differently from that of systems in single particle picture . the recent study of fractional charge is such an example . it is suggested that the noise measurement in kondo regime can lead to non - integer effective charge@xcite . moreover , the recent experimental measurement reveals surprisingly high current noise of carbon nanotube in the kondo regime where the conductance is close to the landauer conductance quantum.@xcite this is in great contrast to the predictions based on the noninteracting picture where the noise is expected to be negligible . in this study , we show another example where the noise behavior of ac - driving quantum dots in kondo regime differs drastically from that of noninteracting dots . for non - interacting quantum conductors , lesovik and levitov ( ll ) have predicted that the noise should show singularities at integers of @xmath2 , where @xmath3 is the dc bias , @xmath4 is the ac frequency.@xcite soon , this prediction was experimentally verified by measuring the noise power spectrum of a tunnel junction driven by microwave field . @xcite further theoretical studies , which in principle are still in single - particle pictures , showed that the singularity is robust against disorder effect and remains distinct at weak ac field@xcite ( actually , the singularity also depends on the ratio of @xmath5 where @xmath6 is the ac strength ) . none of these studies included correlation effect due to the coulomb interaction in the conductor . when there is finite coulomb interaction in the conductor , the noise behavior is expected to show different behavior since the electrons are correlated and the single - particle picture is no - longer valid . _ studied the noise in an ac - driving luttinger liquid.@xcite their results showed that the sharp features in @xmath7 is smoothed by the presence of coulomb interaction . since the electron transport in kondo regime is dominated by the strongly correlated kondo resonance , an immediate question one might ask is what happens to the photon - assisted noise singularities in the kondo regime . there exists some literatures which studied the transport properties of the ac kondo regime.@xcite most of them focus on the conductance or density of states of an ac - driving kondo dot . there have hitherto been few attempts to study the noise properties in ac - driving kondo dot@xcite . recently , the noise of kondo dot with ac gate modulation was studied@xcite by present authors by combining the floquet theorem and the slave boson mean field approximation . this mean - field approximation is valid when the kondo temperature defines largest energy scale . the numerical results showed that the photon - assisted noise of kondo dot shows no singularity at zero temperature and very low bias voltage . this smooth noise behavior was in great contrast to the predictions in single particle picture . limited by the validity of slave - boson mean field approximation , it remains unclear whether such singularity can appear in the kondo regime when the kondo temperature does not define the largest energy scale . it is the purpose of the present paper to investigate the photon - assisted noise of kondo dot beyond the slave boson mean field approximation . our approach is based on the noncrossing approximation ( nca)@xcite . for @xmath8 , where @xmath0 is the coulomb interaction strength , the nca has been shown to give reliable results both in the temperature regimes above and below @xmath1 . the splitting of the kondo resonance by external dc bias can also be captured . however , it fails to reproduce the fermi liquid behavior at very low temperature @xmath9 and gives non - physical results in the mixed - valence region.@xcite a suitable remedy of this shortcoming is possible as shown by kroha _ _ et al.__@xcite . the nca has also been applied to the time - dependent nonequilibrium transport through a kondo dot.@xcite solution of the the time - dependent nca in the time domain has been proposed and implemented.@xcite in the present study , we avoid to go to the pathological regime of nca@xcite and make use of the periodicity of the ac field . we combine the floquet theorem@xcite and the infinite-@xmath0 nca to investigate the transport properties , in particular the current noise properties of quantum dot with ac gate modulation . a suitable floquet - green s function formalism is developed to describe the dynamics of the system . the floquet theorem can capture the multi - photon process in a coherent non - perturbative way . our results show that for low ac frequency and low ac strength , the local density of states ( ldos ) of the kondo dot is almost insensitive to the ac gate field . for high ac frequency and strength , the kondo peak can be significantly suppressed with sidebands appearing in the density of states with distances to the kondo peak of multi - photon energy . such suppression is attributed to the decoherence induced by the photon - assisted processes@xcite . interestingly , the noise of quantum dots in ac - kondo regime shows dramatically different behavior from the ac driving noninteracting dots . the noise as a function of dc bias voltage does not show singularities in the kondo regime as we change the ac parameters from weak to strong values . one may thus utilize the photon - assisted noise behavior to tell apart the single - particle noninteracting resonant transport from the strongly correlated kondo resonance . the present paper is organized as following . in sec . ii , we give the details of the model hamiltonian and the floquet - green s function formalism with nca . in sec . numerical results of the ldos and noise properties of ac - driving kondo dot are presented . for the sake of comparison , results for non - interacting dot are given to show the drastic different noise behaviors with or without many - body correlations . in sec . iv a conclusion is presented . the quantum transport through quantum dot is widely described by the single impurity anderson model coupled to two ideal leads . the system can be driven out of equilibrium by both dc bias and ac gate field . here , the ac field can be realized via a nearby ac gate which modulates the dot energy level . to investigate the coulomb effect on the noise behavior , we are interested in two situations , i.e. noninteracting dot without coulomb interaction and the kondo dot with strong on - site coulomb interaction . for noninteracting dot , the system can be modeled by the single impurity anderson model with zero coulomb strength ( @xmath10 ) coupled to two ideal leads . the hamiltonian can be written as @xmath11 where @xmath12 is the creation operator of @xmath13-electron with spin @xmath14 in the @xmath15th lead ( @xmath16 ) , @xmath17 is the lead electron energy . we assume a symmetric voltage drop across the dot . @xmath18 denotes the coupling matrix element of the lead electron and that in the dot . @xmath19 is the creation operator of electron in the dot with spin @xmath14 . its energy level can be modulated by an external gate voltage as @xmath20 . here spin - degeneracy is implied . usually , the simple assumption of zero coulomb interaction can not correctly capture the electron dynamics in nanostructures . for quantum dot with small - size , the charging energy can be much larger than the other energy scales . in these situation , we can approximately take the limit @xmath8 . for infinite @xmath0 , the dot state can contain either zero or one electron . double occupancy may be neglected . a reliable tool to give a description of quantum dots with level deep below the fermi energy is the slave boson nca@xcite . the slave boson technique introduces some auxiliary particles to eliminate the non - quadratic term in the hamiltonian by replacing the electron operators by @xmath21 and @xmath22 , where @xmath23 annihilates the empty state and @xmath24 destroys a spin @xmath14 single occupation state . the infinite @xmath0 hamiltonian in slave boson language reads @xmath25 since the double occupancy is forbidden for @xmath26 , these slave particle operators must satisfy the constraint condition @xmath27 to study the transport properties out of equilibrium , we make use of the nonequilibrium green s function ( negf ) method . the physical information is usually stored in two propagators , i.e. the retarded , @xmath28 , and the lesser , @xmath29 , green s functions which are defined in the real time axis . these green s functions can be related to the contour - ordered green s function @xmath30 , along the keldysh contour , by the langreth rules.@xcite the retarded green s function can be found from the dyson equation in the time - domain as @xmath31 the lesser green s function is given by the keldysh equation @xmath32 where @xmath33 is the advanced green s function and @xmath34 is the double - time retarded / lesser self - energy . usually , we need another green s function , the greater green s function @xmath35 which can be found by either using the keldysh equation by replacing @xmath36 with the greater self - energy @xmath37 or from the identity : @xmath38 for general time - dependent hamiltonian , the solution of the above equations remains challenging . however , since our hamiltonian is periodic in time , the floquet theorem can be used to simplify the solution.@xcite it is more desirable to work in an enlarged hilbert space which is defined by the direct product of the local basis in space and the basis for the time - periodic functions with the frequency @xmath4 . the basis set of time periodic functions which can be chosen to have the form @xmath39 with @xmath40 gives the so - called floquet basis to take into account the periodicity in a nonperturbative way . we will combine the floquet theory and the green s functions to investigate the time - dependent transport properties of quantum dot systems . the following notations are defined and will be used in the rest of the paper . the fourier transform of a double time function @xmath41 is defined as @xmath42 the periodicity of the function @xmath43 indicates that it can be decomposed into the floquet basis as @xmath44 where @xmath45 is the fourier coefficient of the function @xmath46 . several authors have studied the time - dependent transport by the fourier coefficients of the floquet - green s functions @xcite , where only single label of the floquet states is needed . however , as we will see later , it is more convenient to write this time - period function @xmath47 in a matrix form @xmath48 in floquet space . in the following , all operators expressed in matrix form in floquet space will be denoted by calligraphic symbols . the matrix element @xmath49 are related to the fourier coefficient by @xmath50 in the matrix form of floquet - green s function , it is more convenient to work with the floquet hamiltonian which is defined by @xmath51 using the previously introduced notations , the double - time retarded green s function eq.([grt ] ) can be rewritten in floquet space in a compact form as the resolvent of the floquet - hamiltonian @xcite as @xmath52 where @xmath53 is the double - time retarded self - energy in floquet space and @xmath54 is the floquet - hamiltonian of the isolated dot . the derivation of the above equation is given in the appendix . similarly , the advanced floquet - green s function is given by @xmath55^\dag.\end{aligned}\ ] ] after fourier transform , the lesser floquet - green s function which in time - domain is given by eq . ( [ glt ] ) can be rewritten in a compact form as @xmath56 with @xmath57 represents the matrix form of the lesser self - energy in floquet space . we can see that in the matrix form in floquet space , both the retarded and lesser green s functions ( eq . ( [ fgr ] ) and eq . ( [ fgl ] ) ) have similar structures with the keldysh equations for the green s function in _ stationary _ situations@xcite , though these green s function in time - domain are double - time functions . the main difference is that the floquet - green s function is expressed in an enlarged hilbert space due to the periodicity condition . for noninteracting dot , i.e. @xmath10 , the hamiltonian of the system can be given with quadratic terms only . we assume the energy level of the dot is modulated by a harmonic field . the dot hamiltonian takes the time - dependence @xmath58 . the corresponded floquet hamiltonian is then given by @xmath59_{k , k'}=\{\epsilon_d+k\hbar\omega\}\delta_{k , k ' } + \frac{1}{2}v_g \delta_{k , k'\pm 1}.\end{aligned}\ ] ] in the presence of ac gate , the self - energies are due to the coupling between the dot and the leads . for the dot - lead coupling , the retarded and lesser self - energy which is given in time - domain as @xmath60 where @xmath61 represents the green s function of the lead electrons at equilibrium and @xmath15 is the lead label . the retarded and lesser green s functions of lead electrons are given , respectively , by @xmath62,\end{aligned}\ ] ] @xmath63,\end{aligned}\ ] ] where @xmath64 is the fermi - distribution function in the @xmath15 lead . after fourier transform and some algebra , the self - energies ( eq . ( [ set ] ) ) can be rewritten in the floquet basis as @xmath65 once we neglect the energy shift by the dot - lead coupling , the static self - energy @xmath66 can be given in the so - called wide - band approximation by @xmath67 where @xmath68 . the transport properties can be given by the green s functions and their self - energies of the quantum dot . the expression for the time - dependent current operator from left lead is related to the time derivation of the total number operator in the left lead : @xmath69,\end{aligned}\ ] ] with @xmath70 . in terms of the nonequilibrium green s functions , the expectation value of current can be found by @xmath71\},\end{aligned}\ ] ] where @xmath72 is the advanced self - energy due to the coupling to the left lead . due to the discrete nature of electrons , the current fluctuation is nonzero even at zero temperature . this fluctuation , known as current noise , contains information of electron correlations which is beyond the reach of traditional conductance measurement . this current noise is defined as@xcite @xmath73 where @xmath74 represents the fluctuation of the current operator from the left lead from its expectation value . using the the floquet - green s functions , the time - dependent current can be rewritten in a compact form with similar structure of the time - independent meir - wingreen current formula as @xmath75_{k,0}(\omega ) e^{-ik\omega t}\},\end{aligned}\ ] ] where @xmath76 and @xmath77 are the advanced and lesser self - energy in floquet space due to the coupling to the left leads . to find the time - averaged current over one period , one can simply set @xmath78 in the above formula . inserting the current operator in the definition of the noise definition and after some algebra , the time - averaged current fluctuation at zero frequency can be given in the matrix form of the floquet - green s functions and self - energies as @xmath79_{0,0}.\end{aligned}\ ] ] the above noise formula is exact for noninteracting dot where the hamiltonian of the system is quadratic so that the wick theorem can be applied . in the presence of finite coulomb interaction , we have quartic terms . a direct application of wick theorem is no longer possible . a full diagrammatic expansion is needed to reach the desired formula . however , such task remains formidable since multi - particle green s functions are required . in order to make use of the power of wick theorem , one has to rely on approximations such as the mean field approximation or slave boson techniques to rewrite the hamiltonian in quadratic form . for infinite-@xmath0 , a well established tool to study the transport properties in the kondo regime is the nca@xcite . the nca is the lowest order conserving approximation . it is widely used to investigate the properties of kondo dot at equilibrium or out of equilibrium . using nca , the self - energies of the pseudo - particles are given in the time - domain as @xmath80 where @xmath81 is the greater green s function of electrons in leads . @xmath82,\end{aligned}\ ] ] @xmath83 is the self - energy of the pseudo - fermion while @xmath84 represents the self - energy for the pseudo - boson green s functions . the double time green s functions for the slave particles are defined by : @xmath85 the retarded and lesser green s functions of the pseudo - boson and pseudo - fermion , together with their self - energies are obtained from a set of self - consistent equations . the dyson equations for the retarded and lesser green s functions in time - domain are given by@xcite : @xmath86 @xmath87 @xmath88;\end{aligned}\ ] ] @xmath89.\end{aligned}\ ] ] the greater and advanced green s functions are given in nca by : @xmath90;\\ b^>(t , t')&&=i[b^r(t , t')-b^a(t , t')];\\ d^a(t , t')&&=[d^r(t',t)]^;\\ b^a(t , t')&&=[b^r(t',t)]^.\end{aligned}\ ] ] making use of the periodicity of our problem , the above equations can again be rewritten in matrix form in an enlarged space . the floquet - green s function of the pseudo - fermion and pseudo - boson can be given in a compact form in matrix form as : @xmath91 where @xmath92 and @xmath93 are the floquet hamiltonian of the pseudo - fermion and pseudo - boson , respectively . @xmath94 ( @xmath95 ) is the floquet - green s function of the pseudo - fermion ( pseudo - boson ) in the matrix form . their corresponding self - energies in the matrix form are denoted as @xmath96 and @xmath97 , respectively . the matrix element of the self - energies for pseudo - particles within nca is given in floquet space by : @xmath98 these equations can be solved in a self - consistent way . at each iteration in the numerical calculations , the constraint condition eq . [ cost ] is numerically checked as @xmath99=1.\end{aligned}\ ] ] from the time - derivation of the total electron number in the left lead , the time - averaged current formula can be found with the help of floquet - green s functions as @xmath100(\omega)\}_{0,0}.\end{aligned}\ ] ] due to the strong correlations , an exact formula for noise in kondo regime is formidable , if it exists . one has to make some approximation . along the line for the derivation of eq . ( [ sn ] ) , a noise expression in the slave particle floquet - green s functions within nca can be obtained as @xmath101_{0,0},\end{aligned}\ ] ] in the above derivations , we have , following meir _ _ et al.__@xcite , neglected the vertex correction when decoupling the correlation functions . one merit of this approximation is that it can recover the zero - voltage fluctuation - dissipation theorem@xcite . this merit has been tested in our numerical results . in the following , we will use the above noise formula to investigate the noise properties of ac driving kondo dot . in this section , we apply the floquet - green s function approach with infinite-@xmath0 nca to calculate the ldos and noise properties of ac - driving kondo dot . in the following calculations , we take the lorentzian shape ldos of the lead . the coupling between the dot and lead is then given by @xmath102 where @xmath103 is the half bandwidth of the lead , @xmath104 is the chemical potential . we will take @xmath105 as the energy unit in the following discussions . for the sake of simplicity , we use @xmath106 . throughout calculation , we have fixed @xmath107 without other statement . the half bandwidth of the lead is @xmath108 . the kondo temperature for infinite @xmath0 can be estimated from @xmath109 without other statement which is below the estimated kondo temperature and not in the pathology regime of nca . one fingerprint of the kondo effect is the sharp kondo peak of the ldos at the fermi energy . in the following , we study the time - averaged ldos of quantum dot which is driven by an ac gate voltage . the ldos of the kondo dot is related to the retarded green s function of the dot electrons which can be found from @xmath110 at equilibrium , the retarded green s function after fourier transform is independent of time . when the system is driving by ac field , the retarded green s function after fourier transform is time - dependent and can be given in floquet space as @xmath111_{k_1,k_2}=i\int\frac{d\omega'}{2\pi}[\mathcal{d}^r_{\sigma;k_1 , k'}(\omega+\omega')\mathcal{b}^<_{k',k_2}(\omega'-k_2\hbar\omega)+\mathcal{d}_{\sigma;k_1,k'}^<(\omega+\omega ' ) \mathcal{b}^a_{k',k_2}(\omega'-k_2\hbar\omega)].\end{aligned}\ ] ] the time - averaged ldos can then be easily obtained by @xmath112_{0,0}.\end{aligned}\ ] ] in fig . 1 , we show the calculated ldos @xmath113 as a function of energy @xmath114 of the quantum dot modulated by ac gate via @xmath115 . different ac frequencies are used in our calculated as indicated in the figure . the ratio between the ac strength and ac frequency is fixed at @xmath116 . from fig . 1 , we can see that the ac gate voltage can modify the time - averaged ldos . however , the kondo peak is not suppressed monotonically by increasing ac frequency ( ac strength ) . the evolution of the kondo peak with increasing ac field is enlarged in the inset of fig . 1 . for low ac field ( ac frequency and ac strength ) , for example , @xmath117 , the time - averaged ldos is almost identical with the equilibrium ldos . the ldos against ac field remains robust in the numerical results as up to an external ac frequency of @xmath118 . this value is much larger than the width of the kondo peak which can be estimated by its kondo temperature @xmath1 . from fig . 1 , by increasing the ac frequency from @xmath119 to @xmath120 , one can find that the broad peak around @xmath121 is only slightly lowered while the kondo peak is almost unchanged . for these low frequencies , the time - averaged ldos largely resembles the averaged ldos for quantum dot at equilibrium with time - dependent energy level over a period of the ac modulation . the observed robust of ldos against not - strong ac field agrees with the findings in ref . @xcite where the kondo dot is found to be not sensitive to the weak ac field . however , when the ac field is strong , the adiabatic picture is no longer valid . significant suppression of the kondo peak is observed in fig . 1 for @xmath122 and @xmath123 . at the same time , the width of the kondo peak is broadened . for very large ac field ( for @xmath124 and @xmath125 in fig . 1 ) , the numerical results clearly show the inelastic photon - assisted processes . one can observe in fig . 1 that replica of kondo peak appears in the ldos with a distance to the kondo peak equals the ac frequency . for a non - interacting quantum dot , analysis based on the scattering approach has shown that the photo - assisted noise as a function of the dc bias displays cusps at integer @xmath126@xcite . the derivation of the noise then gives staircase behavior . such singularities can be attributed to the change of the distribution of the transmitted charge by ac field . here , we calculate the photon - assisted noise through a noninteracting quantum dot with ac gate by the floquet - green s function method and noise formula eq . ( [ sn ] ) . identical numerical results are also obtained by the noise formula presented in ref . @xcite which is valid for ac transport in the absence of coulomb interaction . in fig . 2 , we present the time - averaged noise power ( in unit @xmath127 ) as a function of dc bias voltage @xmath3 of ac gate modulated noninteracting dots for different dot energy levels . in the calculation , the temperature is fixed at @xmath128 . the ac frequency is chosen as @xmath129 and the ac strength is @xmath130 . since the magnitude of noise varies a lot with the energy level position , we have rescaled the @xmath131 axis to make the time - averaged noise @xmath132 behavior at @xmath133 more clearly . we can see from fig . 2 that the appearance of cusp in the photon - assisted noise depends on the position of the energy level of the quantum dot . when @xmath134 , i.e. the dot is in resonance , the noise shows an obvious cusp at @xmath135 . the cusp is ascribed to the non - adiabatic photon - assisted tunneling@xcite . since we have chosen a nonzero temperature , the sharpness of this cusp has been smoothed . when we lower the energy level and tune the dot out of resonance , the noise curves become smooth around @xmath136 . these results indicate that the singularity of @xmath7 for noninteracting dot in the presence of ac gate modulation is most pronounce when the dot is situated at resonance . however , when the dot is far from resonance , the noise behaviors become smooth and it is hard for experiments to detect the noise singularity behavior . we have shown that the non - interacting _ resonant _ dot can display noise singularity when the dot is modulated by ac gate . an intuitive analogous between the noninteracting resonant dot and the kondo resonance , where both resonances can reach the conductance quanta ( @xmath137 with spin degeneracy ) , may lead to the naive conclusion that the noise of kondo dot will show singularity in the presence of ac gate since there is kondo resonance at fermi energy . however , this statement is not correct , at least in the strongly correlated infinite-@xmath0 regime , as shown in our numerical results fig . 3 . in fig . 3 , we show the noise properties of the kondo dot with different ac frequency as a function of the applied dc voltage . since we are interested in the noise behavior at @xmath133 , the @xmath138 axis is rescaled by the ac frequency to have a better view at @xmath133 ( note different values of @xmath4 in different curves ) . in fig . 3 , no singularity is observed at @xmath133 for all the ac parameters . for lowest ac frequency ( @xmath139 ) , this is in agree with our previous results obtained by slave - boson mean - field approximation@xcite . the noise at weak ac field is almost identical to the numerical results without ac field . it is interesting to see that both the slave - boson mean - field approximation and noncrossing approximation , which are valid in their respective parameter space , show that the noise of kondo dot remains almost unaffected by the ac gate . the kondo dot behaves as if the ac gate is effectively screened to induce singularity in the noise behavior . comparing with the noninteracting resonance tunneling , the disappear of singularity behavior in noise of kondo dot may be ascribed to the fact that the electrons through the kondo resonance are not _ directly _ modulated by the ac field . although the energy level of the quantum dot @xmath121 is periodically modulated by the ac gate , the kondo peak is not directly driven by the ac field . for weak ac field where the ac frequency or ac strength are not in the order of the coupling strength or the energy level , the ac gate voltage is too small to drive the dot out of the kondo regime . strong correlation can still give rise to the sharp kondo peak at fermi energy . from the numerical results of ldos , we can see that ac modulation of kondo peak is much smaller than the applied ac gate which modulated the energy level of the dot . when the electrons are tunneling though the dot via the kondo resonance , their transport behavior is not significantly modulated by the ac field . the influence of the time - periodicity of the kondo peak to the electrons transport through the dot is negligibly small to show any singularities in the noise behavior . however , when the ac strength is strong enough to be comparable to @xmath121 or @xmath140 , for example @xmath141 in fig . 1 and 3 , the quantum dot can sometimes be driven out of the kondo regime by the ac field . the magnitude of the kondo peak can then be significantly suppressed due to the strong ac field . as a consequence , the conductance declines drastically and deviates from the unitary limit . in these situations , the kondo peak is then significantly suppressed and electron transmission probability becomes too small to show significant noise singularity . since the cusp of noise behavior can be smoothed by finite temperature , it is important to exclude the possibility that the smooth noise behavior shown in fig . 3 is due to the finite temperature ( @xmath142 ) in our numerical simulation . in fig . 4 , we plot the noise behavior as a function of dc bias voltage for the dot with different temperature . the ac frequency and ac strength is fixed at @xmath143 . the temperatures are chosen to be 0.01 , 0.005 and 0.001 , respectively . from fig . 4 , we can see that even for the lowest temperature @xmath144 which is much lower than @xmath1 , no cusp is observed at @xmath133 in the noise behavior . together with the numerical results presented in ref . where no singularity is observed at zero temperature , the disappearance of cups in noise behavior can not be ascribed to the finite temperature . it is a generic feature of the strongly - correlated nature of kondo peak . our numerical results together with those of ref . @xcite which are obtained with the help of the slave - boson technique have clearly shown that the strongly - correlated kondo resonance can display different noise behavior as compared with the noninteracting dot . usually , both the noninteracting resonant tunneling and kondo resonance will lead to similar results in conductance measurement . therefore , it is possible to make use of the distinct photon - assisted noise behavior to tell apart whether the resonant transport is via noninteracting resonance or strongly - correlated kondo resonance . such measurement has been in the reach of present technology . the recent experiment work@xcite has shown the possibility to conduct the noise measurement of nanostructures with ac field . the ability to distinguish the origin of the resonant transport may be helpful to answer the controversy about origin of the 0.7 anomaly in quantum point contacts , where the zero bias anomaly is due to whether the formation of strongly correlated kondo peak@xcite or single - particle resonant peak@xcite is still under debating . we wish the experiment of the photon - assisted noise measurement can be used to reveal the underlying physics of 0.7 anomaly . in conclusion , we have presented a floquet - green s function formalism to investigate the noise properties of quantum dots in the ac - kondo regime . the coulomb correlation is taken into account by the infinite-@xmath0 nca . in principle , the ac effect can be considered non - perturbatively via the infinite floquet states . our results show that the kondo peak is robust against weak ac gate modulation . significant suppression of the kondo peak can be observed when the ac gate field becomes strong . the photon - assisted noise of kondo dots as a function of dc voltage does not show singularities which are expected for noninteracting resonant quantum dot at integer @xmath2 , where @xmath3 is the dc bias voltage and @xmath4 is the ac frequency . these results suggest that we can tell resonant transport apart from the kondo resonance by photon - assisted noise measurement , which can not be distinguished via the conductance measurement . in this appendix , we outline the derivation to reach the resolvent form of the floquet - green s function . it resembles much to the resolvent form of the green s function at steady state . the starting point is the definition of green s function in time domain @xmath145 1 a. c. hewson , _ the kondo problem to heavy fermions _ ( cambridge university press , cambridge , 1993 ) ; t. k. ng and p. a. lee , phys . rev . lett . * 61 * , 1768 ( 1988 ) ; l. i. glazman and m. e. raikh , pisma zh . . fiz . * 47 * , 378 ( 1988 ) [ jetp lett . 47 , 452 ( 1988 ) ] ; d. goldhaber - gordon , h. shtrikman , d. mahalu , d. abusch - magder , u. meirav , and m. a. kastner , nature * 391 * , 1569 ( 1998 ) ; s. . cronenwett , t. h. oosterkamp , and l. p. kouwenhoven , sience * 281 * , 540 ( 1998 ) ; j. nygard , d. h. cobden , and p. e. lindelof , nature * 408 * , 342 ( 2000 ) ; j. park , a. n. pasupathy , j. i. goldsmith , c. chang , y. yaish , j. r. petta , m. rinkoski , j. p. sethna , h. d. abruna , p. l. mceuen , and d. c. ralph , nature * 417 * , 722 ( 2002 ) ; w. liang , m. p. shores , j. r. long , and h. park , nature * 417 * , 725 ( 2002 ) ; v. kashcheyevs , a. aharony , and o. entin - wohlman , phys . rev . b * 73 * , 125338 ( 2006 ) ; n. e. bickers , d. l. cox , and j. w. wilkins , phys . rev . b * 36 * , 2036 ( 1987 ) ; n. e. bickers , rev . mod . phys . * 59 * , 845 ( 1987 ) ; n. s. wingreen , and y. meir , phys . b * 49 * , 11040 ( 1994 ) ; for a review , see r. bulla , t. a. costi , and t. pruschke , rev . 80 * , 395 ( 2008 ) ; ya . m. blantter and m. bttiker , phys . rep . * 336 * , 1 ( 2000 ) ; y. meir and a. golub , phys . lett . * 88 * , 116802 ( 2002 ) ; m. sindel , w. hofstetter , j. von delft , and m. kindermann , phys , rev . * 94 * , 196602 ( 2005 ) ; e. sela , y. oreg , f. von oppen , and j. koch , phys . rev . lett . * 97 * , 086601 ( 2006 ) ; a. golub , phys . b * 75 * , 155313 ( 2007 ) ; o. zarchin , m. zaffalon , m. heiblum , d. mahalu , and v. umansky , phys . rev . b * 77 * , 241303(r ) ( 2008 ) ; t. delattre , c. feuillet - palma , l. g. herrmann , p. morfin , j .- berroir , g. fve , b. plaais , d. c. glattli , m .- s . choi , c. mora and t. kontos , nature physics * 5 * , 208 ( 2009 ) ; g. b. lesovik , and l. s. levitov , phys . lett . * 72 * , 538 ( 1994 ) ; r. j. schoelkopf , a. a. kozhevnikov , d. e. prober , and m. j. rooks , phys . lett . * 80 * , 2437 ( 1998 ) ; a. lamacraft , phys . rev . 91 * , 036804 ( 2003 ) ; m. guigou , a. popoff , t. martin , and a. crpieux , phys . rev . b * 76 * , 045104 ( 2007 ) ; m. h. hettler and h. schoeller , phys . rev . lett . * 74 * , 4907 ( 1995 ) ; t. k. ng , phys . rev . lett . * 76 * , 487 ( 1996 ) ; g .- h . ding and t. k. ng , phys . rev . b * 56 * , 15521(r ) ( 1997 ) ; p. nordlander , m. pustilnik , y. meir , n. s. wingreen , and d. c. langreth , phys . rev . lett . * 83 * , 808 ( 1999 ) ; p. nordlander , n. s. wingreen , y. meir , and d. c. langreth , phys . b * 61 * , 2146 ( 2000 ) ; r. lpez , r. aguado , g. platero , and c. tejedor , phys . rev . lett . * 81 * , 4688 ( 1998 ) ; y. goldin , and y. avishai , phys . rev . lett . * 81 * , 5394 ( 1998 ) ; q. chen and h. k. zhao , europhys . lett . * 82 * , 68004 ( 2008 ) ; b. h. wu , and j. c. cao , phys . rev . b * 77 * , 233307 ( 2008 ) ; h. shao , d. c. langreth , and p. nordlander , phys . rev . b * 49 * , 13929 ( 1994 ) ; r. aguado , and d. c. langreth , phys . rev . b * 67 * , 245307 ( 2003 ) ; j. kroha , and p. wlfle , j. phys . soc * 74 * , 16 ( 2005 ) ; j. kroha , p. wlfle , and t. a. costi , phys . rev . lett . * 79 * , 261 ( 1997 ) ; g. platero and r. aguado , phys . rep . * 395 * , 1 ( 2004 ) ; s. kohler , j. lehmann , and p. hnggi , phys . rep . * 406 * , 379 ( 2005 ) ; a. kaminski , y. v. nazarov , and l. i. glazman , phys . * 83 * , 384 ( 1999 ) ; h. haug and a .- jauho , _ quantum kinetics in transport and optics of semiconductors _ ( springer - verlag , berlin , 1998 ) . j. h. shirley , phys . rev . * 138 * , b979 , ( 1965 ) . m. grifoni and p. hnggi , phys . rep . * 304 * , 229 ( 1998 ) . s. camalet , j. lehmann , s. kohler , and p. hnggi , phys . lett . * 90 * , 210602 ( 2003 ) ; l. e. f. foa torres , phys . b * 72 * , 245339 ( 2005 ) ; l. arrachea , phys . rev . b * 72 * , 125349 ( 2005 ) ; l. arrachea , a. l. yeyati , and a. martin - rodero , phys . rev . b * 77 * , 165326 ( 2008 ) ; b. h. wu and j. c. cao , j. phys.:condens . matter * 20 * , 085224 ( 2008 ) ; a .- jauho , n. s. wingree , and y. meir , phys . b * 50 * , 5528 ( 1994 ) ; j. gabelli , and b. reulet , phys . rev . lett . * 100 * , 026601 ( 2008 ) ; y. meir , k. hirose , and n. s. wingreen , phys . lett . * 89 * , 196802 ( 2002 ) ; a. lassl , p. schlagheck , and k. richter , phys . b * 75 * , 045346 ( 2007 ) ; t .- m . chen , a. c. graham , m. pepper , i. farrer , and d. a. ritchie , phys . b * 79 * , 153303 ( 2009 ) ;	the transport properties of an ac - driving quantum dot in the kondo regime are studied by the floquet - green s function method with slave - boson infinite-@xmath0 noncrossing approximation . our results show that the kondo peak of the local density of states is robust against weak ac gate modulation . significant suppression of the kondo peak can be observed when the ac gate field becomes strong . the photon - assisted noise of kondo resonance as a function of dc voltage does not show singularities which are expected for noninteracting resonant quantum dot . these findings suggest that one may make use of the photon - assisted noise measurement to tell apart whether the resonant transport is via noninteracting resonance or strongly - correlated kondo resonance .
recently , a lot of attentions are focused on the study of the azimuthal anisotropy of the particle production in heavy ion collisions at the relativistic heavy ion collider ( rhic ) and the large hadron collider ( lhc ) . this anisotropy is usually expanded into a fourier series : @xmath12 with @xmath0 and @xmath13 represent the magnitude and direction of @xmath14-order anisotropy , respectively . at low @xmath11 , @xmath0 is driven by the anisotropic pressure gradient associated with the initial spatial asymmetries , with more particles emitted in the direction of largest gradients @xcite . asymmetries giving rise to non - zero @xmath0 are associated with either average shape ( for @xmath6 ) or shapes arising from spatial fluctuations of the participating nucleons @xcite . they can be characterized by a set of multi - pole components ( also known as `` eccentricities '' ) at different angular scale , calculated from the participating nucleons at @xmath15 @xcite : @xmath16 the orientations of the minor axis for each moment @xmath17 , also known as the participant plane ( pp ) are given by @xmath18 when fluctuations are small and linearized hydrodynamics is applicable , each moment of the flow @xmath0 is expected to be independently driven by @xmath19 along @xmath20 @xcite . this may not be true when the fluctuations are large , as the non - linear effects may lead to significant mixing between harmonic flow of different order @xcite . in this paper , they are assumed to be the same to facilitate the study of the correlations between @xmath13 of different physics origins . at high @xmath11 ( @xmath21 ) gev , the @xmath0 is understood to be driven by the path - length dependent energy loss of jets traversing the medium , with more particles emitted along the direction of shortest path - length ( or direction of smallest jet attenuation ) @xcite . in contrast to flow which is sensitive to the global geometry manifested through the global evolution of the created matter , this anisotropy is more sensitive to the local density experienced by the jets as they traverse the matter . nevertheless , since both flow and jet quenching are influenced by the same geometry , the directions of largest pressure gradient for flow and direction of smallest jet attenuation are strongly correlated . in fact , they are often implicitly assumed to be the same in many theoretical calculations @xcite . an explicit study of the correlation between the two directions can help clarifying this assumption . in this paper , we estimate the high-@xmath11 anisotropy coefficients @xmath0 and their associated directions @xmath22 ( qp stands for `` quenching plane '' ) using a simple jet absorption model with event by event fluctuating glauber geometry . we study the correlations between @xmath23 and @xmath22 , and discuss implications of these correlations for the interpretation of the `` ridge '' phenomena in two - particle correlations ( 2pc ) . we use a simple jet absorption / glauber model of ref . @xcite to calculate @xmath0 and @xmath22 . this model has been used previously to study the centrality and path - length dependence of single particle suppression @xmath24 , dihadron suppression @xmath25 and @xmath26 . back - to - back jet pairs are generated according to the binary collision density profile ( @xmath27 ) in the transverse ( @xmath28 ) plane with random orientation . they are then propagated through the medium , whose density is given by the participant density profile ( @xmath29 ) . both profiles are generated with a monte carlo glauber model with event by event fluctuation of positions of nucleons in au ions @xcite . the nucleons are sampled from a woods - saxon distribution with a radius of 6.38 fm and diffuseness of 0.535 fm , with a nucleon - nucleon cross - section of @xmath30 mb . in order to have smooth distributions for @xmath27 and @xmath29 , the nucleons are assumed to have a gaussian profile in transverse plane with a width of @xmath31 fm in @xmath32 and @xmath33 direction similar to ref @xcite . the value of @xmath34 is varied from 0.2 - 0.4 fm , and the nucleon is also assumed to be a uniform disk with a radius of @xmath35 fm . however the final results are found to be insensitive to the details of the nucleon shape , except in peripheral collisions . the jet quenching is implemented via exponential attenuation @xmath36 , where the matter integral @xmath37 is calculated as @xmath38 for jet generated at @xmath39 and propagated along direction @xmath40 . they corresponds to @xmath41 dependence of absorption ( @xmath42 ) in a longitudinal expanding or 1 + 1d medium ( @xmath43 ) with a thermalization time of @xmath44 . the @xmath45 is fixed to 0 by default , but we have checked the @xmath0 do not change much for @xmath46 fm @xcite . the two cases , @xmath47 and @xmath48 , are motivated for the @xmath49 dependence expected for radiative and ads / cft energy loss in 1 + 1d medium @xcite , respectively . the absorption coefficient @xmath50 controls the jet quenching strength and is the only parameter in this calculation . it is tuned to reproduce @xmath51 for 0 - 5% @xmath52 data at rhic after averaging over many glauber events @xcite . this leads to a value of @xmath53 @xmath54 and 0.0968 @xmath55 for @xmath47 and 2 , respectively . figure [ fig:1 ] summarize the basic information obtained from this procedure for one typical au - au event in 0 - 5% centrality interval . panels ( a ) and ( b ) shows the density profile for @xmath29 and @xmath27 , respectively . panel ( c ) shows the normalized probability distribution of @xmath56 : @xmath57 , which is obtained by calculating @xmath56 over all possible di - jet production point @xmath27 and jet propagation direction @xmath58 . this distribution exhibit characteristic high density and low density regions in @xmath59 space , presumably reflecting spatial correlation between the @xmath27 and @xmath29 profiles . panel ( d ) shows the normalized probability distribution of the attenuation @xmath60 . panel ( e ) shows the @xmath61 averaged along the y - axis in panel ( d ) as a function @xmath58 , which is precisely the azimuthal angle dependent suppression @xmath62 . a clear anti - correlation can be seen between the peak magnitude of the @xmath62 and breadth of the @xmath56 distribution in panel ( c ) . this distribution can also be obtained by randomly generating many di - jet pairs according the @xmath27 and propagating them through @xmath29 via eq . we expand it into a fourier series : @xmath63 where @xmath64 represents the average suppression , @xmath65 and @xmath4 represent the magnitude and direction of @xmath3-order harmonic of emission probability distribution , respectively . similar studies of @xmath62 were pursued before in ref . @xcite for a pqcd energy loss in a event by event hydrodynamic underlying event . however it focused primarily on the influence of fluctuations on the event - averaged @xmath62 distribution relative to the @xmath66-order event plane ( ep ) . figure [ fig:1 ] ( f ) shows a distribution calculated from @xmath19 and @xmath4 : @xmath67 it visualizes the shape of the initial geometry that is transformed into the final momentum anisotropy via either flow or jet quenching . a good alignment is seen between @xmath9 and @xmath4 for @xmath68 . it also shows that the large @xmath19 for @xmath2 are strongly damped after jet absorption , leading to very small values of @xmath69 for @xmath2 . the study shown in fig . [ fig:1 ] can be repeated for many events . we divide the simulation data into 5% centrality intervals , each containing about 2500 events . figure [ fig:2 ] shows the distribution of @xmath70 for two centrality intervals . strong positive correlations are obtained for @xmath1 , 2 and 3 , while the correlations are rather weak or even become negative for @xmath2 in eq . is calculated with @xmath71 weighting . we have also repeated the study using @xmath72 weighting for @xmath73 and @xmath74 weighing for @xmath1 @xcite , but very little differences are seen . ] . in heavy ion collisions at rhic and lhc , the @xmath0 is usually measured from particle distribution relative to @xmath75 via eq . [ eq:1 ] @xcite . however it has also been derived from the fourier coefficients of two - particle correlation in relative azimuthal angle @xmath76 @xcite : @xmath77 with @xmath78 the fact that the quenching plane and participant plane do not align exactly with each other implies that the @xmath0 measured relative to @xmath23 is not the same as those contributing to the 2pc in eq . [ eq:6a ] . in other words , it is possible that the @xmath0 obtained from single particle analysis is only a fraction of the true anisotropy resulting from jet quenching : @xmath79 since what is measured in experiment is the event plane not the pp , it is important to check whether the event plane align with qp or not , for example in a hydrodynamic model calculation . figure [ fig:3 ] ( a ) and ( c ) summarize the centrality dependence of @xmath65 for @xmath16 and for @xmath56 and @xmath80 , respectively . significant @xmath69 signals are observed for @xmath8 , while higher - order @xmath69 are usually smaller than 1% . the @xmath81 and @xmath82@xmath83 all show strong centrality dependence , while the @xmath84 and @xmath85 show little centrality dependence for @xmath86 . interestingly , the value of the @xmath84 is consistently larger than that for @xmath85 , and it even exceeds @xmath81 value in most central collisions . this behavior suggest that the path - length dependence of energy loss and initial dipole asymmetry from fluctuations corroborate to produce a large @xmath84 . this large @xmath84 is expect to contribute to the high-@xmath11 @xmath87 signal observed by the atlas collaboration @xcite . figure [ fig:3 ] also shows that the @xmath80 type of path - length dependence induces significantly larger @xmath69 than that for @xmath56 : the increase is almost a factor of two for @xmath1 and @xmath88 . this is also observed in other studies before @xcite . figure [ fig:3 ] ( b ) and ( d ) summarize the centrality dependence of @xmath89 for @xmath90 and for @xmath56 and @xmath80 , respectively . as indicated by eq . [ eq:7 ] , this represents the reduction factor of the @xmath0 when it is measured relative to the @xmath9 . the reduction is small for @xmath6 , except in central collisions where it reaches 15% for @xmath56 and 30% for @xmath80 . however the reduction is significantly larger for @xmath1 and 3 , reaching about 50% for @xmath1 in mid - central collisions . the @xmath89 value becomes negative for @xmath2 in central collisions , reflecting an anti - correlation between @xmath9 and @xmath4 ( already shown in fig . [ fig:2 ] ) . interestingly , @xmath89 values for @xmath1 are always smaller than that for @xmath88 ( more misalignment ) , while @xmath84 is always larger than @xmath85 . the dispersion between the @xmath4 and @xmath9 has important implications on the factorization relation eq . [ eq:6b ] . the factorization of @xmath10 into @xmath0 is obviously valid for correlations between two low @xmath11 particles ( soft - soft correlation ) as both are modulated around @xmath9 . the factorization should also be valid for correlation between a low-@xmath11 particle and a high-@xmath11 particle ( soft - hard correlation ) since it involves the projection of the @xmath0 onto @xmath9 , i.e. @xmath91 . experimental data indeed support this @xcite . however the correlation between two high-@xmath11 particles from two independent hard - scattering processes ( hard - hard correlation ) is expected to be larger than the product of the two single particle @xmath0 : @xmath92 therefore , the factorization can not work simultaneously for soft - soft , soft - hard and hard - hard correlations . the large anisotropy coefficients @xmath69 also has important consequences for the `` ridge '' observed in two - particle correlations @xcite . this `` ridge '' is understood to be the result of the constructive contribution of harmonics at @xmath93 . in the literature , it is referred to as either the `` soft - ridge '' @xcite for soft - soft correlation or `` hard - ridge '' @xcite for soft - hard correlation , respectively . here we show that the correlation between two independently produced high-@xmath11 jets can also produce the `` ridge''-like structure . this `` hard - hard ridge '' can be calculated on a probability basis event - by - event by simply self - convoluting the @xmath62 distribution like fig . [ fig:1 ] ( e ) . examples of these structures are shown in fig . [ fig:4 ] for two representative events in both 0 - 5% and 20 - 25% centrality intervals . the magnitude of the ridge , as well as the away - side shape changes dramatically from event to event . they also changes a lot between the @xmath56 and @xmath80 types of path - length dependence jet absorption . figure [ fig:5 ] show the long - range structures ( solid lines ) obtained from the jet absorption model , averaged over many events . the ridge magnitude increases with centrality to about 1.5% ( 4% ) for @xmath56 ( @xmath80 ) path - length dependence in mid - central collisions . this signal should be measurable with the large statistics dataset from lhc . the dashed lines in fig . [ fig:5 ] show the 2pc predicted from the @xmath0 measured relative to @xmath9 . clearly the misalignments between @xmath4 and @xmath9 reduces the ridge magnitude . the reduction is almost 50% in most central collisions , but decrease to about 20% in mid - central collisions . this suggests the difference between the measured ridge and those predicted by the event plane method could be large and measurable . the anisotropy of high-@xmath11 particle is studied in a simple jet absorption framework with event by event fluctuating geometry . the harmonic coefficients @xmath94 are found to be significant for @xmath95 ( @xmath96 ) but become very small for @xmath2 . the correlation between the quenching plane and participant plane are studied . a strong de - correlation is found for @xmath6 in central collisions and for @xmath1 and @xmath97 over the full centrality range . the correlations become negative for @xmath2 in central collisions . this de - correlation , if also confirmed between the event plane and the quenching plane ( e.g via hydrodynamic model that has dijets embeded ) , is expected to break the global factorization of the two particle fourier coefficient @xmath10 into the @xmath0 for the two single particles . it would also imply that the high-@xmath11 @xmath0 measured relative to the event plane could be significantly smaller than the true anisotropy from path - length dependent jet energy loss . these jet quenching @xmath0 also give rise to long range `` ridge '' structure in two - particle correlations . the predicted ridge amplitude is on the order of 0.5 - 4% depending on the centrality and functional form of the @xmath49 dependence of the energy loss , and should be measurable at the lhc using the correlations between two high-@xmath11 particles with a large rapidity separation . our study bear some similarities to ref . @xcite . however , ref . @xcite uses a cumulant expansion framework instead of monte carlo glauber model for initial geometry , and that it focus on the soft - hard ridge instead of the hard - hard ridge in our case . 29 j. y. ollitrault , phys . d * 46 * , 229 ( 1992 ) . b. alver and g. roland , phys . c * 81 * , 054905 ( 2010 ) [ erratum - ibid . c * 82 * , 039903 ( 2010 ) ] . b. h. alver , c. gombeaud , m. luzum and j. -y . ollitrault , phys . c * 82 * , 034913 ( 2010 ) . p. staig and e. shuryak , phys . c * 84 * , 034908 ( 2011 ) . d. teaney and l. yan , phys . rev . c * 83 * , 064904 ( 2011 ) . z. qiu and u. w. heinz , phys . c * 84 * , 024911 ( 2011 ) . m. gyulassy , i. vitev and x. n. wang , phys . lett . * 86 * , 2537 ( 2001 ) . a. adare _ et al . _ [ phenix collaboration ] , phys . lett . * 105 * , 142301 ( 2010 ) . s. a. bass , c. gale , a. majumder , c. nonaka , g. -y . qin , t. renk and j. ruppert , phys . c * 79 * , 024901 ( 2009 ) . j. jia and r. wei , phys . c * 82 * , 024902 ( 2010 ) . j. jia , w. a. horowitz and j. liao , phys . c * 84 * , 034904 ( 2011 ) . t. renk , h. holopainen , j. auvinen and k. j. eskola , arxiv:1105.2647 [ hep - ph ] . b. betz , m. gyulassy and g. torrieri , phys . c * 84 * , 024913 ( 2011 ) . x. zhang and j. liao , arxiv:1202.1047 [ nucl - th ] , j. liao and e. shuryak , phys . * 102 * , 202302 ( 2009 ) . a. drees , h. feng and j. jia , phys . c * 71 * , 034909 ( 2005 ) . b. alver , m. baker , c. loizides and p. steinberg , `` the phobos glauber monte carlo , '' arxiv:0805.4411 [ nucl - ex ] . m. chesler , k. jensen , a. karch and l. g. yaffe , phys . d * 79 * , 125015 ( 2009 ) . c. marquet and t. renk , phys . b * 685 * , 270 ( 2010 ) . a. adare _ et al . _ , phys . * 101 * , 232301 ( 2008 ) . a. m. poskanzer and s. a. voloshin , phys . c * 58 * , 1671 ( 1998 ) . atlas collaboration , arxiv:1203.3087 [ hep - ex ] . star collaboration , phys . c * 80 * , 064912 ( 2009 ) . cms collaboration , arxiv:1201.3158 [ nucl - ex ] . m. daugherity [ star collaboration ] , j. phys . g g * 35 * , 104090 ( 2008 ) . s. gavin , l. mclerran and g. moschelli , phys . c * 79 * , 051902 ( 2009 ) [ arxiv:0806.4718 [ nucl - th ] ] . g. moschelli and s. gavin , nucl . a * 830 * , 623c ( 2009 ) [ arxiv:0907.4362 [ nucl - th ] ] .	the azimuthal anisotropy due to path - length dependent jet energy loss is studied in a simple jet absorption model that include event by event fluctuating glauber geometry . significant anisotropy coefficients @xmath0 are observed for @xmath1,2 and 3 , but they are very small for @xmath2 . these coefficients are expected to result in a `` ridge '' for correlations between two independently produced jets . the correlations between the orientation of the @xmath3-order anisotropy induced by jet absorption ( @xmath4 ) and the @xmath3-order participant plane ( @xmath5 ) responsible for harmonic flow are studied . tight correlations are observed for @xmath6 in mid - central collisions , but they weaken significantly for @xmath7 . the correlations are positive for @xmath8 , but become negative in central collisions for @xmath2 . the dispersion between @xmath4 and @xmath9 is expect to break the factorization of the fourier coefficients from two - particle correlation @xmath10 into the single particle @xmath0 , and has important implications for the high-@xmath11 ridge phenomena .
let @xmath3 be a complete @xmath4dimensional riemannian manifold , @xmath5 . let @xmath6 and @xmath7 @xmath8 , be bounded , simply connected domains of class @xmath9 of @xmath3 such that @xmath10 and @xmath11 if @xmath12 . consider the multiply connected domain @xmath13 and set @xmath14 , @xmath15 . with this notation _ _ @xmath16 let @xmath17 be given . in this paper we investigate the dirichlet problem@xmath18{l}q_{h}\left ( u\right ) { \small : = \operatorname{div}}\frac{\nabla u}{\sqrt{1+\left\vert \nabla u\right\vert ^{2}}}+nh{\small = 0}\text { in } \omega,\text { } u\in c^{2,\alpha}\left ( \overline{\omega}\right ) \\ u\|_{\gamma}=0,\text { } u\|_{\gamma_{i}}=h\text { , } i=1, ... ,m \end{array } \right . \label{dirh}\ ] ] where @xmath19 and @xmath20 are the divergence and the gradient in @xmath3 . if @xmath21 is a solution of ( [ dirh ] ) then the graph of @xmath21 is a compact constant mean curvature @xmath22 hypersurface of @xmath0 oriented by a unit normal vector @xmath23 such that @xmath24 and whose boundary lies in the slices @xmath25 . the dirichlet problem ( [ dirh ] ) was studied in the work @xcite for @xmath26 , in @xcite when @xmath27 and @xmath28 has only two connected components and in @xcite when @xmath29 and @xmath30 . in these works existence results are obtained when the height @xmath31 is less than or equal to a constant which depends on the mean curvature of @xmath28 , the distance between the connected component of @xmath28 , the dimension @xmath32 , and on the diameter of @xmath33 and @xmath22 if @xmath34 . in @xcite some nonexistence results also are established . in the case @xmath30 we observe that theorem 1 of @xcite , an extension of the classical jenkins - serrin result - theorem 2 of @xcite - , gives us an existence result with the upper bound of @xmath31 depending on @xmath32 , on the mean curvature and on the injectivity radius of @xmath28 . the main motivation to study the problem ( [ dirh ] ) is that , for @xmath34 , we did not find in the literature , even for @xmath35 , a result where the hypersurface @xmath36 is not assumed to be mean convex . we explore this situation when @xmath3 is a hadamard manifold ( theorem [ h+ ] ) . we observe that in @xcite the authors conclude , for @xmath35 , the existence and uniqueness of @xmath22-graphs for a large class of prescribed boundary data over @xmath36 where @xmath36 is not necessarily mean convex but , however , @xmath37 with @xmath33 simply connected . relatively to the minimal case , we obtain theorem [ minimal ] , whose estimate on @xmath31 is more in line with that in theorem 2.1 of @xcite than that in theorem 1 of @xcite . despite being difficult to say if our result improves or not the estimate of @xmath31 given in @xcite in general , for some domains @xmath33 we got some gain ( see remark [ r1 ] ) . an extra motivation to our work is the problem proposed by a. ros and h. rosenberg in remark 4 of @xcite : given two convex jordan curves in distinct parallel planes of @xmath38 , is there a cmc annulus having such curves as boundary ? besides of the aforementioned works , this situation was also investigated in @xcite , @xcite and @xcite ( for some characterization results , see @xcite and @xcite ) . the results obtained so far do not give a complete answer to this problem . our results give some contribution in the @xmath0 context . we fix some notations : the mean curvature of @xmath28 with respect to the unit normal vector field @xmath39 to @xmath28 pointing to @xmath33 will be denoted by @xmath40 and @xmath41 let @xmath42 be the riemannian distance in @xmath3 . denote by @xmath43 , the biggest positive numbers such that the exponential maps @xmath44 and @xmath45 are diffeomorphisms ( here , @xmath46 ) and set @xmath47 and @xmath48 domains . ] denote by @xmath49 ( @xmath50 ) the solution of the equation @xmath51 we prove : [ minimal]let @xmath3 be a complete @xmath4dimensional riemannian manifold , @xmath5 . if@xmath52 where @xmath53 then the dirichlet problem ( [ dirh ] ) has a unique solution for @xmath30 . [ h+ ] assume that @xmath3 is a hadamard manifold ( complete , simply connected riemannian manifold with nonpositive sectional curvature ) and that @xmath33 is contained in a geodesic ball of radius @xmath54 of @xmath3 . then there is a positive constant @xmath55 , explicitly given in ( [ tc ] ) such that , given @xmath56 we have@xmath57 \sigma)}{\delta\left [ \left ( n-1\right ) \left\vert h_{\inf}^{\partial \omega}\right\vert + nh\left ( 1+\delta\right ) \right ] } -\frac{h\re^{2}}{1+\sqrt{1-h^{2}\re^{2}}}>0,\ ] ] where @xmath58 } \right\ } , \ ] ] and the dirichlet problem ( [ dirh ] ) has a unique solution if @xmath59 . we shall use the continuity method from elliptic pde theory in the proof of the main results . then , we need to construct local barriers relatively to the dirichlet problem ( [ dirh ] ) ( see @xcite , p. 333 ) . we work with the construction of the local barriers relatively to the points in @xmath36 and in @xmath60 at same time , using @xmath61 to means both @xmath62 and @xmath63 and having in mind that , for @xmath64 , @xmath65 if @xmath66 and @xmath67 if @xmath68 . we consider , at @xmath64 , an orthonormal referential frame @xmath69 _ _ _ _ of _ _ _ _ @xmath70 _ _ _ _ @xmath71 where _ _ _ _ @xmath72 [ lcg]given @xmath73 , set @xmath74 , @xmath75 and consider @xmath76 given by @xmath77 suppose @xmath78 . then @xmath79 in @xmath61 _ _ if@xmath80 ^{3}\right ) \delta d+nh\left ( 1+\left [ \psi^{\prime}\right ] ^{2}\right ) ^{3/2}\leq0 , \label{eg}\ ] ] where @xmath81 is the laplacian in @xmath3 . we have @xmath82 . after some calculus we see that @xmath83 is @xmath84 notice that @xmath85 and so@xmath86{c}2\psi^{\prime}\psi^{\prime\prime}\text { if } i = n\\ 0\text { if } i\neq n \end{array } \right . , \label{eiw2}\]]@xmath87{c}\psi^{\prime}\text { if } i = n\\ 0\text { if } i\neq n \end{array } \right . \label{eiw}\ ] ] and@xmath88{c}\psi^{\prime\prime}e_{n}+\psi^{\prime}\nabla_{e_{n}}^{e_{n}}\text { if } i = n\\ \psi^{\prime}\nabla_{e_{i}}^{e_{n}}\text { if } i\neq n \end{array } \right . . \label{dgwei}\ ] ] it follows from ( [ dgwei ] ) that@xmath89 and from ( [ eiw2 ] ) and ( [ eiw ] ) that @xmath90 ^{2}\psi^{\prime\prime}. \label{s2}\ ] ] plugging ( [ s1 ] ) and ( [ s2 ] ) in ( [ q0w ] ) , as @xmath91 ^{2}$ ] , we obtain @xmath92 ^{2}\right ) ^{-3/2}\left [ \psi^{\prime\prime}+\left ( \psi^{\prime}+\left [ \psi^{\prime } \right ] ^{3}\right ) \delta d\right]\ ] ] and , therefore , @xmath79 if @xmath80 ^{3}\right ) \delta d+nh\left ( 1+\left [ \psi^{\prime}\right ] ^{2}\right ) ^{3/2}\leq0.\ ] ] [ lcl]if @xmath93 @xmath94 , then @xmath95 in @xmath61 . since @xmath96 is compact , there is @xmath97 such that @xmath98 where @xmath39 is the normal unit vector to @xmath28 pointing to @xmath33 . consider @xmath99 \rightarrow\left ( 0,+\infty\right ) $ ] in @xmath100 \right ) $ ] defined by@xmath101 , \ ] ] where @xmath102 is @xmath103 or @xmath104 , according @xmath61 is @xmath62 or @xmath63 . we have @xmath105 . \label{f2f}\ ] ] setting @xmath106 the mean curvature of @xmath107 , where @xmath108 . since @xmath109 and @xmath110 , we have @xmath111 let @xmath112 \longrightarrow u$ ] be the arc - length parametrized geodesic such that @xmath113 and @xmath114 . we have from ( [ ric ] ) and ( [ f2f ] ) that @xmath115 for all @xmath116 $ ] . since @xmath117 is a extension of @xmath39 to @xmath61 and ( [ f1f ] ) , ( [ c2 ] ) occurs , it follows from theorem 5.1 of @xcite that@xmath118 then@xmath119 \leq\left ( n-1\right ) c. \label{estimlapla}\ ] ] [ propos]suppose @xmath93 @xmath94 . given @xmath120 and @xmath78 set @xmath121 and define @xmath122 by@xmath123 } \left [ \cosh^{-1}\left ( \alpha+\lambda\left [ \mu+nh\left ( 1+\lambda\right ) \right ] s\right ) -\cosh^{-1}\left ( \alpha\right ) \right ] \text { , } \label{carapsi}\ ] ] @xmath74 , @xmath64 . then @xmath124 as in ( [ w ] ) satisfies @xmath79 in @xmath125 where @xmath126 from lemma [ lcg ] and lemma [ lcl ] we have @xmath127 in @xmath61 if@xmath128 \left ( \psi^{\prime}\right ) ^{3}+nh\left ( \psi^{\prime}\right ) ^{2}+\left [ \left ( n-1\right ) c+nh\right ] \psi^{\prime}+nh\leq0 \label{o}\ ] ] where @xmath129 , @xmath74 . notice that ( [ o ] ) can be rewritten as@xmath130 ^{3}\right ) + nh\left ( 1+\left [ \psi^{\prime}\right ] ^{2}\right ) \leq0 , \label{or}\ ] ] where @xmath131 , with @xmath132 . we choose@xmath133 where @xmath134 are positive constants to be determined and @xmath135 . setting @xmath136 , since @xmath137 it follows that@xmath138 and @xmath139 then , from ( [ or ] ) we see that @xmath79 if@xmath140 \psi^{\prime } + nh\left ( u^{2}-1+a^{2}b^{2}\right ) \leq0\ ] ] that is , if@xmath140 ab+nh\left ( u^{2}-1+a^{2}b^{2}\right ) \left ( u^{2}-1\right ) ^{1/2}\leq0.\ ] ] we assume @xmath134 such that @xmath141 . then the last inequality is true if@xmath142 which is true if@xmath143 as @xmath136 , the last inequality is@xmath144 s+b\alpha+nh\alpha^{2}-b\leq0 . \label{orw}\ ] ] let @xmath145 . we assume @xmath146 = b\lambda + nh\lambda^{2}>b\alpha+nh\alpha^{2}.\ ] ] we see that ( [ orw ] ) is true for @xmath147 $ ] , where @xmath148 is given by ( [ eps ] ) . this concludes the proof of the proposition . [ cor_pro]under the hypothesis of proposition [ propos ] , @xmath149 where @xmath49 ( @xmath50 ) is the solution of the equation @xmath150 , @xmath151 . in particular , @xmath152 } { \delta\left ( \mu+nh\left ( 1+\delta\right ) \right ) } \label{ro}\ ] ] where @xmath153 notice that , since @xmath154 , @xmath155 set @xmath156 we have @xmath157 iff@xmath158 the equation ( [ eqdelta ] ) has a unique solution @xmath49 ( @xmath50 ) , and @xmath49 is the maximum ( global ) point for @xmath159 . then @xmath160 the equality ( [ ro ] ) follows immediately from definition of @xmath161 and @xmath148 ( observing that both depend on @xmath162 and @xmath163 ) . [ height]let @xmath3 be a hadamard manifold.let @xmath164 @xmath165 be a compact graph of constant mean curvature @xmath34 over a domain @xmath166 and such that @xmath167 and let @xmath168 be the height of @xmath164 . if @xmath33 is contained in a normal ball in @xmath3 of radius @xmath169 then @xmath170 let @xmath171 the center of the normal ball and set @xmath172 , @xmath173 . consider at @xmath173 an orthonormal referential frame @xmath174 _ _ _ _ of _ _ _ _ @xmath70 _ _ _ _ @xmath175 , where _ _ _ _ @xmath176 . let @xmath177 , where@xmath178 we have @xmath179 given by @xmath180 .\ ] ] as the sectional curvature of @xmath3 is @xmath181 , by the laplacian comparison theorem we have @xmath182 , where @xmath183 is the euclidean distance . then@xmath184 = 0,\ ] ] where the equality is due that , for @xmath35 , the graph of @xmath185 is a spherical cap . the result follow now of the fact that the maximum height of the graph of @xmath186 is @xmath187 proof of theorem [ minimal ] since @xmath188 where @xmath53 from lemma [ cor_pro ] there is @xmath189 , @xmath162 close enough to @xmath190 , such that , setting @xmath191 \text { , } \label{npsi}\ ] ] @xmath74 , the function @xmath192 , @xmath193 constant , satisfies @xmath194 in @xmath195 where @xmath196 moreover , @xmath197 notice that @xmath198{c}u_{\gamma}^{\epsilon}:=\left\ { z\in u_{\gamma};d(z,\gamma)<\varepsilon \right\ } \text { if } u = u_{\gamma}\\ u_{\gamma_{i}}^{\varepsilon}:=\left\ { z\in u_{\gamma_{i}};d(z,\gamma _ { i})<\varepsilon\right\ } \text { if } u = u_{\gamma_{i}}\end{array } \right . , \ ] ] and then we can consider the supersolutions @xmath199 and @xmath200 , @xmath201 $ ] , relatively to the operator @xmath202 , given by @xmath203 and @xmath204 respectively , where @xmath205 is given by ( [ npsi ] ) . it follows that @xmath206 in @xmath207 and @xmath208 in @xmath209 . moreover , @xmath210 in @xmath211 and , then , in @xmath212 . let @xmath213 given by @xmath214{c}0\text { if } p\in\gamma\\ h\text { if } p\in\gamma_{i}\text { , } i=1, ... ,m \end{array } \right . . \label{nphi}\ ] ] from the maximum principle , it follows that _ _ @xmath215 and _ _ @xmath216 are upper local barriers relatively to the boundary data @xmath217 and , for lower local barriers relatively to @xmath217 , just take @xmath218 and @xmath219 in the domains @xmath220 and @xmath221 respectively . now , set@xmath222;\exists u_{t}\in c^{2,\alpha}\left ( \overline{\omega } \right ) \text { satisfying } q_{0}\left ( u_{t}\right ) = 0,\text { } u_{t}\|_{\partial\omega}=t\varphi\}\text{.}\ ] ] we have @xmath223 since @xmath224 . moreover , since @xmath202 is a uniformly elliptic operator on @xmath225 we can apply the implicit function theorem in banach spaces to conclude that @xmath226 is an open . now , we apply a standard sequence of arguments to conclude that @xmath226 is closed . let @xmath227 a sequence with @xmath228 $ ] . for each @xmath32 , let @xmath229 satisfying @xmath230 , @xmath231 . from the barriers above _ _ it follows that the sequence @xmath232 has uniformly bounded @xmath233 norm . moreover@xmath234 it follows of section @xmath235 of @xcite that there is @xmath236 such that @xmath237 and , consequently , @xmath238 with the constant @xmath239 independent of @xmath32 . hlder estimates and pde linear elliptic theory - see @xcite - give us that @xmath240 is equicontinous in the @xmath241 norm for some @xmath242 . it follows that @xmath240 contains a subsequence converging uniformly on the @xmath243 norm to a solution @xmath244 . regularity theory of linear elliptic pde ( @xcite ) implies that @xmath245 . therefore , @xmath226 is closed , that is , @xmath246 $ ] and this gives us the existence result . the uniqueness of the solution is a consequence of the maximum principle for the difference of two solutions . [ r1 ] : denote by @xmath247 the biggest positive number such that the normal exponential map @xmath248 is a diffeomorphism . we call @xmath247 the `` injectivity radius of @xmath28 '' . notice that @xmath249 and we can have @xmath250 . the estimative of @xmath31 in theorem 1 of @xcite , relatively the dirichlet problem ( [ dirh ] ) , is @xmath251 , where @xmath252 and @xmath253 , for some @xmath254 . then , depending of the domain @xmath33 , we have some improvement on the estimate of _ _ @xmath31 in theorem [ minimal ] when we compare with that in theorem 1 of @xcite . for example , if @xmath33 is such that @xmath255 and @xmath256 moreover , using @xmath257 instead of @xmath247 we are more in line with theorem 2.1 of @xcite and theorem 1.1 of @xcite . set @xmath258 , and @xmath259 \sigma)\right ] ^{2}+\mu\delta^{2}\re\left [ \mu\re+2n\left ( 1+\delta\right ) \right ] + \delta n\left ( 1+\delta\right ) \left [ \delta n\left ( 1+\delta\right ) -2\cosh^{-1}(1+\delta\mu\sigma)\right ] } , \label{tc}\ ] ] where@xmath260 } \right\ } .\ ] ] we first notice that @xmath261 in fact , @xmath262 since @xmath263 on the other hand , @xmath264 since we have ( [ m0 ] ) and @xmath265 } \leq\frac{2\delta\mu\cosh^{-1}\delta } { \mu\delta^{2}\re\left [ 2n\left ( 1+\delta\right ) \right ] } = \left ( \frac{1}{\re}\right ) \frac{\cosh^{-1}\delta}{\delta n\left ( 1+\delta\right ) } < \frac{1}{\re}\text{.}\ ] ] let @xmath266 . we claim that @xmath267 \rho)}{\delta\left [ \mu+nh\left ( 1+\delta\right ) \right ] } -\frac{h\re^{2}}{1+\sqrt{1-h^{2}\re^{2}}}>0 , \label{le0}\ ] ] where @xmath268 } \right\ } .\ ] ] indeed , notice that , if @xmath30 then ( [ le0 ] ) is true . suppose @xmath34 . after some calculations , we have ( [ le0 ] ) iff@xmath269 \rho)}{h}-2\delta^{2}\mu nh\re^{2}\left ( 1+\delta\right ) -\delta^{2}n^{2}h^{2}\re^{2}\left ( 1+\delta\right ) ^{2}\nonumber\\ & > \left [ \cosh^{-1}(1+\delta\left [ \mu+nh\left ( 1+\delta\right ) \right ] \rho)\right ] ^{2}+\nonumber\\ & + \re^{2}\delta^{2}\mu^{2}-2\delta n\left ( 1+\delta\right ) \cosh ^{-1}(1+\delta\left [ \mu+nh\left ( 1+\delta\right ) \right ] \rho ) . \label{le1}\ ] ] notice that , since @xmath270 @xmath271 \rho)\label{le4}\\ & \leq\cosh^{-1}(1+\delta\left [ \mu+\frac{n}{\re}\left ( 1+\delta\right ) \right ] \rho)\nonumber\end{aligned}\ ] ] and then , from ( [ le4 ] ) , we have that if @xmath272 \rho)\right ] ^{2}+\re^{2}\delta^{2}\mu^{2}-2\delta n\left ( 1+\delta\right ) \cosh^{-1}(1+\delta\mu\rho)\nonumber\end{aligned}\ ] ] then ( [ le1 ] ) is true . notice that since @xmath273 we have @xmath274 and , then , ( [ le5 ] ) occurs if@xmath275 \sigma)\right ] ^{2}+\re^{2}\delta^{2}\mu^{2}-2\delta n\left ( 1+\delta\right ) \cosh^{-1}(1+\delta\mu\sigma).\nonumber\end{aligned}\ ] ] notice that ( [ le6 ] ) is equivalent to @xmath276 and this concludes the proof of the claim . as @xmath277 , it follows that@xmath278 \rho)}{\delta\left [ \mu+nh\left ( 1+\delta\right ) \right ] } -\frac{h\re^{2}}{1+\sqrt{1-h^{2}\re^{2 } } } , \label{ee}\ ] ] and , as @xmath59 , from lemma [ cor_pro ] there is @xmath189 , @xmath162 close enough to @xmath190 , such that , setting @xmath279 s)-\cosh^{-1}\left ( \alpha\right ) } { \delta\left [ \mu+nh\left ( 1+\delta\right ) \right ] } \text { , } \label{npsi2}\ ] ] @xmath74 , the function @xmath192 , @xmath193 constant , satisfies @xmath280 in @xmath195 , where @xmath281 and , moreover , from ( [ ee ] ) , @xmath282 as @xmath198{c}u_{\gamma}^{\epsilon}:=\left\ { z\in u_{\gamma};d(z,\gamma)<\varepsilon \right\ } \text { if } u = u_{\gamma}\\ u_{\gamma_{i}}^{\epsilon}:=\left\ { z\in u_{\gamma_{i}};d(z,\gamma _ { i})<\varepsilon\right\ } \text { if } u = u_{\gamma_{i}}\end{array } \right . , \ ] ] we can consider the supersolutions @xmath283 and @xmath284 , @xmath285 $ ] , relatively to the operator @xmath286 , given by @xmath287 and @xmath288 respectively , where @xmath205 is given by ( [ npsi2 ] ) . it follows that @xmath206 in @xmath207 and @xmath289 in @xmath290 . moreover , @xmath291 in @xmath292 and @xmath293 in @xmath294 . let @xmath295 as in ( [ nphi ] ) . from lemma [ height ] and from the maximum principle , it follows that _ _ @xmath215 and _ _ @xmath216 are upper local barriers relatively to the boundary data @xmath296 . from the theorem [ minimal ] , since @xmath297 , there is @xmath298 satisfying @xmath299 , @xmath300 , @xmath301 , @xmath8 . as vertical translation in @xmath0 are isometries , we use vertical translations of @xmath302 to obtain lower barrier relatively to the boundary data @xmath217 . now , set@xmath222;\exists u_{t}\in c^{2,\alpha}\left ( \overline{\omega } \right ) \text { satisfying } q_{th}\left ( u_{t}\right ) = 0,\text { } u_{t}\|_{\partial\omega}=t\varphi\}\text{.}\ ] ] we have @xmath223 since @xmath224 . moreover , @xmath226 is open by the implicit function theorem in banach spaces . from the barriers above , we obtain a priori uniform @xmath303 estimates for the family of dirichlet problems @xmath304 @xmath305 , which give us that @xmath226 is closed ( by similar sequence of arguments exposed in the last paragraph of proof of theorem [ minimal ] ) . the uniqueness of the solution is a consequence of the maximum principle for the difference of two solutions . a. aiolfi , j. ripoll , m. soret : the dirichlet problem for the minimal hypersurface equation on arbitrary domains of a riemannian manifold , preprint ( a preliminary version on http://arxiv.org/abs/1307.7884 ) .	we establish existence and uniqueness of compact graphs of constant mean curvature in @xmath0 over bounded multiply connected domains of _ _ @xmath1 with boundary lying in two parallel horizontal slices of @xmath2
the derived category approach to b - type d - branes @xcite appears to be extremely powerful . it allows one to go beyond the picture of d - branes as vector bundles over submanifolds so that @xmath0-corrections can be correctly understood . for example , the fact that b - type d - branes must undergo monodromy as one moves about the moduli space of complexified khler forms can be expressed in the derived category language @xcite . the main purpose of this paper is to try to classify which d - branes can become massless at a given point in the moduli space . again the language of derived categories will be invaluable . in order for an object in the bounded derived category of coherent sheaves to represent a d - brane it must be `` @xmath1-stable '' . criteria for @xmath1-stability have been discussed in @xcite although it is not clear that we yet have a mathematically rigorous algorithm for determining stability . despite this , in simple examples such as in the above references and @xcite one can compute stability with a fair degree of confidence . in particular if you have reason to believe that a certain set of a d - branes are stable at a given point in the moduli space then one can move along a path in moduli space and see how the spectrum of stable states changes . there is considerable evidence @xcite that such changes in @xmath1-stability depend only on the homotopy class of the path in the moduli space of conformal field theories . the fact that changes in @xmath1-stability _ do _ depend on the homotopy class of such paths was used in @xcite to `` derive '' kontsevich s picture of monodromy at least in the case of the quintic threefold . the moduli space of conformal field theories may be compactified by including the `` discriminant locus '' consisting of badly - behaved worldsheet theories . typically one expects such theories to be bad because some d - brane has become massless @xcite . indeed , the monodromy seen in @xcite around parts of this discriminant locus was intimately associated to massless d - branes . it is this link between massless d - branes and monodromy that we wish to study more deeply in this paper . in simple cases as one approaches a point in the discriminant locus , a single d - brane becomes massless . of more interest to us is the case where an infinite number become massless . in @xcite one of the authors studied components of the discriminant locus corresponding to what was called `` ez - transformations '' . namely if one has a threefold @xmath2 with some complex subspace @xmath3 , there may be a point in khler moduli space where @xmath3 collapses to a complex subspace @xmath4 of lower dimension than @xmath3 . we will see that it is then _ the derived category of @xmath4 _ that describes the massless d - branes associated to this transformation . a particular autoequivalence was naturally associated to a particular ez - transformation and it was conjectured in @xcite that such an autoequivalence resulted from the associated monodromy . we will call this conjecture the `` ez - monodromy conjecture '' . one purpose of this paper is to justify this conjecture . because of the nature of our understanding of d - branes and string theory it will not be possible to rigorously _ prove _ any hard theorems about d - branes . instead we will have to play with a number of conjectures whose interdependence leads to considerable evidence of the validity of the overall story . in particular , on the one hand we have the ez - monodromy conjecture and , on the other hand , we have our conjecture concerning which d - branes become massless . these two conjectures are interlinked by @xmath1-stability as we discuss in section [ s : mon ] . in particular , in section [ ss : single ] we discuss an older conjecture concerning single massless d - branes . in section [ ss : gen ] we then review a framework for the more general case which is linked to the simpler case in section [ ss : new ] for a particular example . the physical interpretation of the general case is then given in section [ ss : int ] . the link discussed in section [ ss : new ] between the simple case of a single d - brane becoming massless and an infinite number becoming massless depends upon a mathematical result which is derived in section [ s : comp ] . this section is more technical than the other sections and may be omitted by the reader if need be . that said , it shows how the sophisticated methods of derived categories are directly relevant to the physics of d - branes . in section [ s : app ] we discuss a natural hierarchy of cases . the familiar `` conifold''-like situation arises where @xmath4 is a point and only one soliton becomes massless . if @xmath4 has dimension one then the derived category of @xmath4 has more structure . this case corresponds to seiberg witten theory of some nonabelian gauge group . we study an explicit example of this elsewhere @xcite . the case where @xmath4 has complex dimension two is more complicated as the derived category now has a rich structure . we show that it appears to be similar to the spectrum of massless d - branes one gets from a decompactification . we also see that it demonstrates how 2-branes wrapped around a 2-torus can become massless . at first sight this appears to contradict t - duality but we will see that this is not actually the case . finally , for completeness , in section [ ss : exof ] we discuss the case of an exoflop which is awkward to fit into our general classification but still yields a simple result . b - type d - branes on @xmath2 correspond to objects in the bounded derived category of coherent sheaves on @xmath2 @xcite . a given object @xmath5 is represented by a complex . we may then construct another object @xmath6 $ ] by shifting this complex @xmath7 places to the left . such a shift or `` translation '' is a global symmetry of physics if it is applied simultaneously to all objects @xcite . relative shifts are significant an open string stretched between @xmath5 and @xmath8 is not equivalent to an open string stretched between @xmath6 $ ] and @xmath8 if @xmath9 . we would like to consider the case of moving to a point in moduli space where a single physical d - brane @xmath5 becomes massless . because of the global shift symmetry all of its translates @xmath6 $ ] are equally massless . thus an infinite number of objects in @xmath10 are becoming massless even though only one d - brane counts towards any physical effects of this masslessness as it would be computed by strominger @xcite for example . the analysis of @xmath1-stability in @xcite showed that monodromy is intimately associated to massless d - branes . this should not be surprising since monodromy can only occur around the discriminant and the discriminant is associated with singularities in the conformal field theory associated with massless solitons @xcite . consider an oriented open string @xmath11 stretched between two d - branes in a threefold @xmath2 . in the derived category language this is written as a morphism between two objects in @xmath10 @xmath12 these two objects may or may not form a bound state according to the mass of the open string @xmath11 . if @xmath11 is tachyonic then we have a bound state la sen @xcite . ( as we emphasize shortly @xmath5 is really an anti - brane in such a bound state . ) a real number is defined modulo some integer such as 6 @xcite . periodicity can also appear , if desired , in floer cohomology ( see @xcite for example ) which is supposedly mirror to the structure we are considering . for simplicity we ignore such a possibility . to take such an effect into account one should probably quotient the derived category by such translations . ] ( dubbed a `` grade '' in @xcite ) @xmath13 is associated to each stable d - brane . we assume @xmath13 varies continuously over the moduli space and is defined mod 2 by the central charge @xmath4 : @xmath14 the precise definition of @xmath13 is discussed at length in @xcite . in @xcite it was argued that the mass squared of the open string in ( [ eq : fab ] ) is then proportional to @xmath15 allowing the stability of this bound state to be determined . one of the key features of the derived category which makes it so useful for the study of solitons is the way that bound states are described using _ distinguished triangles_. the open string @xmath11 between @xmath5 and @xmath8 is best represented in the context of a distinguished triangle @xmath16\|{[1]}&\\ \mathsf{a}\ar[rr]^f&&\mathsf{b}.\ar[ul ] } \label{eq : abc}\ ] ] the `` @xmath17 $ ] '' represents the fact that one must shift one place left when performing the corresponding map . the object @xmath18 , which is equivalent to the `` mapping cone '' @xmath19 , is then potentially a bound state of @xmath20 $ ] and @xmath8 . as explained in @xcite , @xmath21 $ ] should be thought of as an anti-@xmath5 . the triangle also tells us that @xmath8 is potentially a bound state of @xmath5 and @xmath18 . equally @xmath5 is a bound state of @xmath8 and @xmath22 $ ] . the `` @xmath17 $ ] '' could be interpreted as keeping track of which brane should be treated as an anti - brane . the fact that @xmath10 copes so well with anti - branes demonstrates its power to analyze d - branes . the other approach , namely k - theory , should be considered the derived category s weaker cousin since it only knows about d - brane _ charge _ ! now suppose that @xmath5 is stable and becomes massless at a particular point @xmath23 in the moduli space . furthermore , let us assume that the _ only _ massless d - branes at @xmath23 are of the form @xmath24 $ ] for any @xmath25 . let us take a generic complex plane with polar coordinates @xmath26 passing through @xmath23 at the origin and assume that @xmath27 behaves as @xmath28 near @xmath23 for @xmath29 some real and positive constant . that is , we assume that @xmath27 has a simple zero at @xmath23 . suppose @xmath8 does not have vanishing mass . it follows that @xmath30 and @xmath31 are equal at @xmath23 and nonzero . in particular if we circle the point @xmath23 by varying @xmath32 , these central charges will be constant close to @xmath23 . furthermore , if @xmath18 can be a marginally bound state of anti-@xmath5 and @xmath8 near @xmath23 , then , according to the rules of @xcite , we have @xmath33 near @xmath23 . this allows us to rewrite ( [ eq : abc ] ) including the differences in the @xmath13 s for the open strings ( i.e. , sides of the triangle ) to give @xmath16_(0.4){1+a - b+\frac{\theta}{\pi}}\|{[1]}&\\ \mathsf{a}\ar[rr]^f_{b - a-\frac{\theta}{\pi } } & & \mathsf{b},\ar[ul]_0 } \label{eq : abcg}\ ] ] where @xmath34 and @xmath35 at @xmath36 . the stability of a given vertex of this triangle depends upon the number on the opposite side being less than 1 . by `` stability '' we mean relative to this triangle only . a given d - brane may decay by other channels . it follows that @xmath18 becomes stable for @xmath37 while @xmath8 becomes unstable for @xmath38 . note that @xmath5 is always stable near @xmath23 consistent with our assumptions . based on this idea that we `` gain '' @xmath18 and `` lose '' @xmath8 as @xmath32 increases , we can try to formulate a picture for monodromy around @xmath23 . the meaning of monodromy is that after traversing this loop in the moduli space we should be able to relabel the d - branes in such a way as to restore the physics we had before we traversed the loop . it is important to note that monodromy is not really the statement that a certain d - brane manifestly `` becomes '' another d - brane explicitly as we move through the moduli space . it is much more accurately described as a relabeling process . since stability is a physical quality , we are forced to relabel @xmath8 since it has decayed . the obvious candidate in the above case is to call it @xmath18 . thus monodromy would transform @xmath8 into @xmath18 . life can be more complicated than this however . if we have an open string @xmath39 , then , since @xmath33 when @xmath8 decays to @xmath40 $ ] , @xmath18 will immediately decay further to @xmath41 plus another @xmath20 $ ] . suppose @xmath5 is `` spherical '' in the sense of @xcite which means @xmath42)=\c$ ] for @xmath43 or 3 , and @xmath42)=0 $ ] otherwise . this condition is always satisfied in the context of this subsection i.e. , only @xmath5 and its translates become massless . a long exact sequence associated to ( [ eq : abc ] ) then implies @xmath44 it follows that this second decay will occur if @xmath45 . iterating this process one sees that @xmath8 will decay splitting off an @xmath20 $ ] a total of @xmath46 times . finally we should also worry about homomorphisms between @xmath8 and @xmath24 $ ] for other values of @xmath25 . we refer to the example in section 4 of @xcite for a detailed example of exactly how this happens in a fairly nontrivial example . all said , allowing for all these decays , @xmath18 becomes a number of @xmath5 s ( probably shifted ) together with @xmath47^{b_1}\oplus \mathsf{a}[-2]^{b_2}\oplus\ldots\right)\to\mathsf{b}\right ) , \label{eq : aaz}\ ] ] where @xmath48,\mathsf{b})\\ & = \dim\ext^n(\mathsf{a},\mathsf{b } ) . \end{split}\ ] ] the cone ( [ eq : aaz ] ) may be written more compactly as @xmath49 where @xmath50 is the complex of @xmath51-vector spaces @xmath52 we refer to @xcite for further explanation of the notation . we can also write more heuristically @xmath53 interpreted navely , we have shown that , upon increasing @xmath32 from @xmath54 to @xmath55 , an object @xmath8 will decay and a canonically associated object @xmath56 will become stable and appear as one of the decay products of @xmath8 . what is desired however is monodromy _ once _ around @xmath23 , i.e. , @xmath32 should only increase by @xmath57 . we will indeed claim that monodromy once around @xmath23 replaces @xmath8 by @xmath56 . for some objects , increasing @xmath32 by only @xmath57 ( at the appropriate starting point ) will cause the complete decay of @xmath8 into @xmath56 . thanks to its rather simple cohomology , this always happens for @xmath58 , the structure sheaf of a point @xmath59 . therefore the relabeling process under monodromy should replace @xmath8 by @xmath56 . there are undoubtedly many other objects @xmath60 under which this increase in @xmath32 by only @xmath57 would not induce the entire decay to @xmath61 . this does nt matter however , we can still leave physics invariant by relabeling @xmath60 by @xmath61 . for example in an extreme case , both @xmath60 and @xmath61 may be stable with respect to the above triangle both before and after increasing @xmath32 by @xmath57 . it is therefore harmless to relabel one of these states as the other . well , it is fine saying that it is harmless to relabel @xmath60 by @xmath61 , but why are we _ forced _ to relabel like this ? the reason is that we know that physics must be completely invariant under monodromy which implies that the relabeling must amount to an _ autoequivalence _ of @xmath10 . one can indeed show that @xmath62 defines an autoequivalence , this transformation can be written as a fourier mukai transform removing this objection . ] of @xmath10 so long as @xmath5 is spherical @xcite . what s more , as argued in @xcite , it is pretty well the _ only _ autoequivalence that works . to be more precise , once we have argued that the specific objects @xmath58 undergo monodromy given by @xmath62 then all other objects must undergo the same monodromy up to some possible multiplication by some fixed line bundle @xmath63 . note that the central charge @xmath4 is also a physical quantity . insisting that monodromy acts correctly in this case amounts to insisting that the d - brane `` charges '' @xmath64 transform under monodromy . this is precisely the same monodromy on @xmath65 that one deduces from mirror symmetry as in @xcite . this determines that the above line bundle @xmath63 is trivial ( in a considerably overdetermined way ! ) . it is known ( see @xcite for example ) that @xmath62 then induces the correct transformation on these charges indeed this was the reason why @xmath62 was conjectured as the monodromy action in the first place @xcite ! it is worth noting that in some special cases the transformation @xmath62 has nothing to do with decay . consider how the spherical object @xmath5 itself transforms : @xmath66 , \end{split } \label{eq : ocon}\ ] ] where we use the convention of @xcite by underlining the zero position when necessary . such a transformation can not be argued from @xmath1-stability however . clearly an open string between @xmath5 and itself ( perhaps translated ) can not have a mass that depends upon some angle as we orbit the conifold point as clearly the mass is constant . instead one could argue that the transform ( [ eq : ocon ] ) occurs simply because @xmath27 has a simple zero at the conifold point and thus @xmath67 shifts by @xmath68 as we loop around the conifold point . then we can apply the rule @xmath69)=\varphi(\mathsf{a})+n$ ] from @xcite .. since monodromy is a relabeling process , one is free to do this , but it looks unnatural from the perspective of associating monodromy with @xmath1-stability . ] we must therefore view ( [ eq : cone1 ] ) as being motivated by @xmath1-stability for _ most _ but not all of the objects in @xmath10 . note that the fact that the obvious physical requirement that monodromy be an autoequivalence of @xmath10 can force ( [ eq : cone1 ] ) to be the required transform for _ all _ the objects in @xmath10 once @xmath1-stability has established it for a few elements . this was the basis of the proof in the case of the quintic in @xcite . anyway , all said we have motivated the following conjecture ( which , in perhaps a slightly different form , is due to kontsevich @xcite , horja @xcite and morrison @xcite ) : if we loop around a component of the discriminant locus associated with a single d - brane @xmath5 ( and thus its translates ) becoming massless then this results in a relabeling of d - branes given by an autoequivalence of the derived category in which @xmath8 becomes @xmath70 . [ conj : a ] this transformation was also motivated by its relation to mirror symmetry and studied at length by seidel and thomas @xcite . it is then natural to ask what happens more generally , i.e. , if more than just a single d - brane @xmath5 becomes massless . in order to answer this we need to set up a general description of how one might analyze monodromy in a multi - dimensional moduli space . there are two paradigms for monodromy both of which are useful : 1 . the discriminant locus decomposes into a sum of irreducible divisors . pick some base point in the moduli space and loop around a component of the discriminant `` close '' to the base point . 2 . restrict attention to a special rational curve @xmath71 in the moduli space . this rational curve contains two `` phase limit points '' , in a sense to be described below , and a single point in the discriminant . the loop in question is around this unique discriminant point . in the case of the one parameter models , such as the quintic , these two paradigms coincide . the moduli space _ is _ @xmath72 and the discriminant locus is a single point . if a component of the discriminant intersects @xmath71 transversely then we can again have agreement between these two pictures of monodromy . in general the discriminant need not intersect @xmath71 transversely a fact we use to our advantage in section [ ss : new ] . we now recall the relationship between the discriminant locus and phases as analyzed in @xcite . the following is a very rapid review . please refer to the references for more details . to make the discussion easier we suffer a little loss of generality and assume we are in the `` batyrev - like '' @xcite case @xmath2 being a hypersurface in a toric variety . the data for @xmath2 is then presented in the form of a point set @xmath73 which is the intersection of some convex polytope with some lattice @xmath74 . see @xcite , for example , for more details of this standard construction . the conformal field theory associated to this data then has a phase structure where each `` phase '' is associated to a regular triangulation of @xmath73 @xcite . the real vector space in which the khler form lives is naturally divided into a `` secondary fan '' of all possible phases . one cone of this fan is the khler cone for @xmath2 where we have the `` '' phase . mirror to @xmath2 , @xmath75 is described as the zero - set of a polynomial @xmath76 in many variables . the points in @xmath73 are associated one - to - one with each monomial in @xmath76 . thus the data @xmath73 is associated to deformations of complex structure of @xmath75 via the monomial - divisor mirror map @xcite . if we model the moduli space of complex structures on @xmath75 by the space of coefficients in @xmath76 , then the discriminant locus can be computed by the failure of @xmath76 to be transversal . this can be mapped back to the space of complexified khler forms on @xmath2 . the result is that part of the discriminant asymptotically lives in each wall dividing adjacent phases in the space of khler forms . that is to say , if we tune the @xmath77-field suitably we can always hit a bad conformal field theory as we pass from one phase to another . thus we may associate singular conformal field theories with phase transitions . the discriminant itself is generically reducible . the combinatorial structure of this reduction has been studied in detail in @xcite . in particular , any time an @xmath25-dimensional face of the convex hull of the set @xmath73 contains more than @xmath78 points , the resulting linear relationship between these points yields a component of @xmath79 . one may then follow an algorithm presented in @xcite to compute the explicit form of each component . the general picture then is of a discriminant with many components with each component having `` fingers '' which separate the phases from each other . each phase transition is associated with fingers from one or more component of @xmath79 . torically each maximal cone in the secondary fan is associated to a point in the moduli space which gives the limit point in the `` deep interior '' of the associated phase . the real codimension - one wall between two maximal cones corresponds to a rational curve @xmath71 passing through two such limit points . the rational curve @xmath71 will intersect the discriminant locus in one point as promised earlier in this section . one component of @xmath79 is distinguished it corresponds to the case of viewing the full convex hull as a face of itself . this is called the `` primary '' component of @xmath79 . closely tied in with conjecture [ conj : a ] ( and at least partially attributed to the same authors ) is the following conjecture at any point on the primary component of @xmath79 ( reached by a suitable path from a suitable basepoint ) the 6-brane associated with the structure sheaf @xmath80 and its translates become massless . at a generic point no other d - branes become massless . [ conj : b ] this idea was perhaps first discussed in @xcite . it is certainly a very natural conjecture the primary component of the discriminant is a universal feature for any manifold and so must be associated with the masslessness of a very basic d - brane . the fact that it works for the quintic was explicitly computed in @xcite , and presumably it is possible to verify the conjecture in a much larger class of examples . we will assume this conjecture to be true . the khler cone is a particular maximal cone in the secondary fan corresponding to the `` '' phase . let us concentrate on the walls of the khler cone . a typical situation as we approach the wall of the khler cone is that an exceptional set @xmath3 collapses to some space @xmath4 . we depict this as @xmath81^q\ar@{^{(}->}[r]^i & x\\ z } \label{eq : ez}\ ] ] where @xmath82 is an inclusion ( which may well be the identity ) and @xmath83 is a fibration with a strict inequality @xmath84 . associated with such a wall in the secondary fan we have a rational curve @xmath71 in the moduli space connecting the large radius limit point with some other limit point . we wish to consider the monodromy associated to circling the point in the discriminant in @xmath71 . the resulting autoequivalence on the derived category has been studied in @xcite where it was dubbed an `` ez - transformation '' . the simplest example would be the case of the quintic threefold which has only one deformation of the khler form . this single component of the khler form gives the overall size of the manifold . thus the `` wall '' ( i.e. , the origin ) corresponds to @xmath2 collapsing to a point . in this case @xmath82 is the identity map and @xmath4 is a point . this is the case discussed above in section [ ss : single ] . indeed , it appears that in all cases where @xmath4 is a point , the resulting monodromy amounts to a transform of the type studied in section [ ss : single ] . it is precisely when @xmath4 is more than just a point the case of interest to us . the precise form of the `` ez - monodromy conjecture '' which associates an autoequivalence of @xmath10 with a given ez - transform was given in @xcite . rather than appealing this conjecture , let us derive the simplest example of a more general case , by assuming the conjectures above , dealing with the case of a single massless d - brane . we will look at the well - known example @xcite , where @xmath2 is a degree 8 hypersurface in the resolution of a weighted projective space @xmath85 . the mirror @xmath75 is then a quotient of the same hypersurface with defining equation @xmath86 the `` algebraic '' coordinates on the moduli space are then given by @xmath87 the primary component of @xmath79 can be computed as @xmath88 the edge of the convex hull containing the points labeled by @xmath89 leads to another component @xmath90 with @xmath91 . @xmath2 can be viewed as a k3-fibration @xmath92 . in this case the component of the khler form given asymptotically ( for @xmath93 ) by @xmath94 controls the size of the k3 fibre . the component of the khler form given asymptotically by @xmath95 gives the size of the @xmath96 base . the base @xmath96 is made very large by setting @xmath97 . in this case , we hit the primary component @xmath98 of the discriminant when @xmath99 . let us refer to this point as @xmath100 . increasing @xmath101 beyond this value moves one out of the phase into the hybrid `` @xmath96-phase '' where the model is best viewed as a fibration with base @xmath96 and a landau ginzburg orbifold as fibre @xcite . fixing @xmath102 and varying @xmath101 spans a rational curve @xmath71 in the moduli space shown in figure [ f : y0 ] . we would like to analyze the monodromy around the singularity @xmath100 in figure [ f : y0 ] . clearly the transition associated with this monodromy consists of collapsing @xmath2 onto the @xmath96 base . in the language of ( [ eq : ez ] ) , @xmath103 , i.e. , the inclusion map @xmath82 is the identity , and @xmath104 . the map @xmath83 is given by the fibration map @xmath105 . in a way , we have constructed the simplest possible example where @xmath4 is more than just a point . now the useful trick is that the monodromy around the singularity in figure [ f : y0 ] can be written in terms of other monodromies that we already understand . this was originally described in @xcite , while this feature was also exploited in @xcite . the full moduli space near @xmath100 is shown in figure [ f : full ] ( with complex dimensions shown as real ) . the rational curve @xmath71 given by @xmath102 corresponds to an infinite radius limit and as such we understand the monodromy around it ( see for example @xcite ) . ( we will follow closely the notation and analysis of @xcite ) . let @xmath106 refer to the autoequivalence of @xmath10 we apply upon looping this curve . it follows that @xmath107 where @xmath108 is the divisor class of a k3 fibre in @xmath2 . meanwhile let @xmath109 refer to the autoequivalence of @xmath10 we apply upon looping the primary component @xmath110 ( i.e. , denote @xmath111 of section [ ss : single ] by @xmath109 ) . then from conjectures [ conj : a ] and [ conj : b ] we know that @xmath112 it follows ( see , for example section 5.1 of @xcite for an essentially identical computation ) that the autoequivalence for the desired loop shown in figure [ f : y0 ] around @xmath100 is given by @xmath113 the desired goal therefore is to find the autoequivalence of @xmath10 obtained by combining the transforms in the above form . the result is that @xmath114 where @xmath115 is an autoequivalence that acts on @xmath10 by @xmath116 section [ s : comp ] is devoted to the proof of this statement . let us review briefly what is exactly meant by the rather concise notation of ( [ eq : picon ] ) . given the map @xmath117 and a sheaf @xmath118 on @xmath2 we may construct the `` push - forward '' sheaf @xmath119 on @xmath4 by associating @xmath120 with @xmath121 for any open set @xmath122 . the @xmath123 appearing in ( [ eq : picon ] ) is the right - derived functor of this push - forward map . this @xmath123 `` knows '' about the cohomology of the fibre of @xmath105 ( see , for example , chapter iii of @xcite ) . the pull - back map @xmath124 is defined for sheaves of @xmath125-modules , and in particular for locally free sheaves , and thus for vector bundles . the map @xmath124 appearing in ( [ eq : picon ] ) is the corresponding left - derived functor . it is a central result of the theory of derived categories @xcite that @xmath124 is the left - adjoint of @xmath123 : @xmath126 for any @xmath127 and @xmath128 . it follows that @xmath129 thus the most natural morphism that would appear in ( [ eq : picon ] ) is the image of the identity on the right - hand side of equation ( [ eq : id1 ] ) under this natural isomorphism . one can show that this is indeed the case . let us interpret ( [ eq : picon ] ) in light of our discussion of monodromy from @xmath1-stability in section [ ss : single ] . to aid our discussion consider how one might rewrite the monodromy result ( [ eq : mo ] ) for the primary component of the discriminant . let @xmath130 be the constant map of @xmath2 to a single point . one can then show , using the fact that sheaf cohomology is equivalent to @xmath131 , which , in turn , is also given by the global section functor @xmath132 @xcite , that ( [ eq : mo ] ) is equivalent to @xmath133 now , the only d - brane ( up to translation in @xmath10 ) which becomes massless in this case is @xmath80 , which is equal to @xmath134 , where we denote the trivial ( very trivial ! ) line bundle on the point @xmath101 as @xmath51 . that is , massless d - branes for the primary component of the discriminant are given by @xmath135(something ) . the @xmath131 in ( [ eq : moc ] ) then gives a natural map to form a cone as required . the expression ( [ eq : heur ] ) immediately dictates that we may interpret @xmath136 in a similar way . _ the d - branes becoming massless at the point @xmath100 in the moduli space correspond to @xmath137 for some @xmath138 . _ the push - forward map @xmath123 can be viewed as the natural ingredient required to form the cone from ( [ eq : id1 ] ) . there is one technical subtlety here which needs to be mentioned . the set of objects of the form @xmath137 for any @xmath139 is not closed under composition by the cone construction . if we write @xmath140 for two objects @xmath141 , then we may only write @xmath142 when there is a relationship between the morphisms @xmath143 . unfortunately such an @xmath144 need not exist for arbitrary @xmath11 . thus we might more properly say that the set of massless d - branes are _ generated _ by objects of the form @xmath137 where we allow for composing such states . flushed with success at interpreting the autoequivalence given by this example we now write down the most obvious generalization for the more general ez - transform ( [ eq : ez ] ) . first of all , we expect @xmath145(something ) to be a massless d - brane on @xmath3 . @xmath3 is mapped into @xmath2 by the inclusion map @xmath82 . the push - forward map @xmath146 is then `` extension by zero '' of a sheaf which is the obvious way of mapping d - branes on @xmath3 into d - branes on @xmath2 . our next conjecture is then any d - brane which becomes massless at a point on a component of the discriminant associated with an ez - transform is generated by objects of the form @xmath147 for @xmath139 . [ c : massless ] this implies a corresponding autoequivalence for the monodromy from @xmath1-stability : @xmath148 for some `` natural '' map @xmath149 . to compute @xmath150 we use the same trick as ( [ eq : id1 ] ) . introduce the functor `` @xmath151 '' as the right - adjoint of @xmath146 : @xmath152 for any @xmath153 and @xmath128 . the existence of @xmath151 is one of the most important features of the derived category in algebraic geometry @xcite and ( [ eq : radj ] ) may be regarded as a generalization of serre duality . now we have @xmath154 this leads to a natural map @xmath155 . this implies the monodromy around the discriminant in the wall associated to a phase transition given by an ez - transform leads to the following autoequivalence on @xmath10 @xmath156 [ c : ez ] this is equivalent to the ez - monodromy conjecture of @xcite . in particular , it was proven there that this is indeed an autoequivalence of @xmath10 . we should perhaps emphasize that we have not _ proven _ conjectures [ c : massless ] and [ c : ez ] . rather we have used the known connection between @xmath1-stability and monodromy and generalized the example we considered in the simplest and most obvious way . note that we may derive conjecture [ conj : a ] from conjectures [ c : massless ] and [ c : ez ] as follows . suppose @xmath4 is a point , then @xmath147 can only be one thing , so we have a single massless d - brane . furthermore , @xmath157 now becomes the cohomology functor giving @xmath158 this also shows that the massless d - brane is @xmath159 i.e. , the d - brane wrapping around @xmath3 as observed in @xcite . for completeness we should also obtain the monodromy as one circles the discriminant in the opposite direction . going back to the triangle ( [ eq : abcg ] ) we see that _ decreasing _ @xmath32 would result in @xmath18 being replaced by @xmath160)[-1]$ ] . this implies we modify the above monodromy arguments to consider the transformation under which @xmath161,\ ] ] for some `` natural '' map @xmath162 . that is , the massless objects bind `` to the right '' of @xmath18 in the mapping cone rather than to the left . the `` @xmath163 $ ] '' is needed because of the asymmetrical definition of the mapping cone we need to keep @xmath18 in its original position . the only nontrivial step in copying the above argument is that to construct @xmath164 we need a left - adjoint functor for @xmath145 . given that @xmath165 we can construct such a functor from @xmath166 thus our desired transform is given by @xmath167.\ ] ] it was shown in @xcite that this is indeed the inverse of the transformation given in conjecture [ c : ez ] . in this section we will prove ( [ eq : pf ] ) . for completeness , we choose to adopt a more general point of view and work with the class of calabi yau fibrations over projective spaces . thus we cover examples such as elliptic fibrations over @xmath168 as discussed in section [ eq : sf ] , as well as the case of k3 fibrations over @xmath96 as desired in section [ ss : new ] . readers not familiar with manipulations in the derived category may well wish to accept the result and skip this section . having said that , some of the methods used in this section are very powerful and may have many other applications to d - brane physics . for the sake of brevity we use the formalism of kernels to describe the fourier - mukai transforms . for the convenience of the reader we review some of the key notions involved . the notations follow those of @xcite . for @xmath2 a non - singular projective variety , an object @xmath169 determines an exact functor of triangulated categories @xmath170 by the formula @xmath171 where @xmath172 is projection on the first factor , while @xmath173 is projection on the second factor . the object @xmath174 is called the _ kernel_. the convenience in using kernels comes about because of the following natural isomorphism of functors @xmath175 the composition of the kernels @xmath176 is defined as @xmath177 where @xmath178 is the obvious projection from @xmath179 to the relevant two factors . there is an identity element for the composition of kernels : @xmath180 where @xmath181 is the diagonal morphism . in this section @xmath2 is assumed to be a smooth calabi yau fibration of dimension @xmath7 over @xmath182 with @xmath117 the fibration map . for us , the calabi yau fibration structure simply means that @xmath117 is a flat morphism ( see section iii.9 of @xcite ) with the generic fibre a calabi yau variety of dimension @xmath183 further assumptions on the calabi yau fibration will be added shortly . in order to set the functors @xmath184 @xmath109 and @xmath185 of the previous section on firm mathematical footing and to define them in the more general context of this section , we need to describe the kernels that induce them as exact functors according to formula ( [ eq : ker ] ) . the following commutative diagram contains most of the maps that we use in the sequel : @xmath186^j & \\ x\ar[d]_\pi & x \times x \ar[l]_{p_1 } \ar[r]^{p_2 } \ar[ld]_{\pi_1 } \ar[dr]^{\pi_2}\ar[d]^{\pi\times\pi } & x\ar[d]^\pi \\ z & z\times z\ar[l]^{s_1 } \ar[r]_{s_2 } & z\ , . } \ ] ] the maps are mostly projections , that are obvious from the context , except for @xmath187 the diagonal of @xmath2 and @xmath117 the fibration map . we now define the fourier mukai functor @xmath188 to be the autoequivalence of @xmath10 induced by the kernel @xmath189 this functor acts on @xmath10 by ( compare to ( [ eq : defl ] ) ) @xmath190 note that the use of the notation @xmath188 for this functor is consistent with the one used in the previous section , since in the case when @xmath2 is a k3 fibration over @xmath191 we have @xmath192 we also define the exact functor @xmath109 induced by the kernel @xmath193 with @xmath194 and @xmath195 the natural restriction map . we can quote for example lemma 3.2 of @xcite to conclude that the action of this functor on @xmath10 is indeed given by ( [ eq : mo ] ) . we make the assumption that the sheaf @xmath80 is spherical ( as defined in section [ ss : single ] ) . this ensures that the functor @xmath109 is an autoequivalence of @xmath196 finally , we define the exact functor @xmath115 to be the so - called fibrewise fourier mukai transform associated to the calabi yau fibration @xmath197 ( see , for example , @xcite , @xcite , @xcite ) . the functor @xmath115 is induced by the kernel @xmath198 with @xmath199 viewed as a sheaf on @xmath200 ( extension by zero ) , and @xmath201 the restriction map . to ensure that the functor @xmath115 is indeed an autoequivalence ( fourier mukai functor ) , we assume that the sheaf @xmath80 is ez spherical . in the language of @xcite , this means that there exists a distinguished triangle in @xmath202 ( @xmath203 ) of the form ) to hold . ] @xmath204 \to \o_z [ 1].\ ] ] note that the sphericity and ez - sphericity conditions are both satisfied in the specific example of a k3 fibration over @xmath96 analyzed in the previous section of this paper . we now justify the use of the notation @xmath115 to denote the fibrewise fourier mukai functor by showing that its action on @xmath10 is indeed , even in the higher dimensional situation , given by the formula ( [ eq : picon ] ) of the previous section . the fibre product @xmath205 fits in the fibre square diagram @xmath206_{q_2}\ar[r]^(0.6){q_1}&x\ar[d]^{\pi}\\ x\ar[r]^{\pi}&z\ , , } \ ] ] and let @xmath207 denote the canonical embedding . since @xmath105 is of finite type and flat , we can apply `` cohomology commutes with the base change '' ( prop . iii 9.3 of @xcite or for a more general form prop . ii 5.12 of @xcite ) : @xmath208 for some @xmath8 in @xmath196 on the other hand , @xmath209 and @xmath210 so we can write @xmath211 note that the last line in the previous formula represents the action on @xmath10 of the exact functor @xmath212 induced as in ( [ eq : ker ] ) by the kernel @xmath199 ( shorthand for @xmath213 ) . this shows that @xmath214 as desired . we are now ready to start discussing the main goal of this section which is to prove the following relation between the defined fourier mukai functors : @xmath215 equivalently , the same formula can be expressed using kernels as @xmath216 where , for any integer @xmath217 @xmath218 of course , the case @xmath219 ( @xmath191 ) of ( [ eq : pfd ] ) is precisely formula ( [ eq : pf ] ) of the previous section . before we move on with the technicalities of the proof , a few remarks are in order . note that the parentheses in the two formulae are simply decorative : the composition of functors , as well as the composition of kernels are associative ( but , of course , not commutative ! ) . for a fixed object @xmath220 in @xmath221 the functors @xmath222 and @xmath223 from @xmath224 to @xmath225 are exact functors between triangulated categories ( i.e. they preserve the distinguished triangles ) . therefore , for any integer @xmath217 we can start with the distinguished triangle defining the kernel @xmath226 @xmath227,\ ] ] and apply to it the operations @xmath228 and @xmath229 from the left , and right , respectively . but @xmath230 and @xmath231 the notation @xmath232 ( the exterior tensor product ) will be used quite often in what follows and simply designates @xmath233 as a shorthand for later convenience we introduce the following objects in @xmath234 @xmath235 and @xmath236 for @xmath237 to justify the definition of the kernel @xmath238 we need to explain how to define the morphism @xmath239 we start with the canonical pairing map on @xmath240 @xmath241 and lift it to @xmath200 @xmath242 we claim that @xmath243 indeed , for the fibre square @xmath244_k\ar[r]^t&z\ar[d]^{\delta_z}\\ x\!\times \!x\ar[r]^{\pi\times\pi}&z\times z\ , . } \ ] ] with @xmath245 the `` diagonal map '' of the fibre square ( [ diag : fq1 ] ) , and @xmath246 flat , we can apply again `` cohomology commutes with the base change '' to obtain @xmath247 . since @xmath248 the previous formula can be written as @xmath249 which is exactly ( [ eq : fibrpr ] ) . the morphism @xmath250 is then defined as the composition @xmath251 note that @xmath252 therefore we have to show that @xmath253 note that the case @xmath254 is immediate since in that case @xmath4 reduces to a point , and the fourier mukai transforms @xmath115 and @xmath255 coincide . an important rle in what follows will be played by beilinson s resolution of the diagonal in @xmath256 @xcite @xmath257 where @xmath258 is the sheaf of holomorphic @xmath82-forms on @xmath259 for any integer @xmath217 @xmath260 we define the complexes @xmath261 on @xmath262 to be the following truncated versions of beilinson s resolution @xmath263 arranged such that the sheaf @xmath264 is located at the 0-th position . define @xmath265 we claim that there exists a natural map is ez spherical , it can be shown that @xmath266 therefore , the described nonzero morphism from @xmath267 to @xmath268 is essentially unique . ] @xmath269 that can be defined at the level of complexes , where , as usual , @xmath268 denotes the complex on @xmath200 with the only non - zero component located at the 0-th position . to see this , we first make the remark that there exists a map of complexes @xmath270 such a map of complexes is well defined , since the complex @xmath261 is a piece of beilinson s resolution . we can now proceed as in ( [ eq : comp ] ) and define the desired morphism in @xmath271 as the composition @xmath272 the key result to be proved in this section is the following : before proving it , let us convince ourselves that the claim implies ( [ eq : aim ] ) . the complex @xmath273 is quasi - isomorphic to the sheaf @xmath274 ( more precisely , to the complex on @xmath262 having the sheaf @xmath274 at the 0-th position ) . therefore , the @xmath275 case of the claim states that @xmath276 since by ( [ eq : fibrpr ] ) @xmath277 we see that the kernel @xmath278 is indeed isomorphic to @xmath279 we now proceed with the inductive proof of the claim . the induction is performed with respect to @xmath280 the case @xmath281 is clear , since @xmath282 and @xmath283 to prove the inductive step @xmath284 we start with the natural maps @xmath285 and @xmath286 applying the _ functors _ @xmath287 and @xmath288 we get two more maps , and we can form the commutative square @xmath289\ar[d ] & { \cal s}_i \ar[d ] \\ { \cal t}_{i+1 } \ar[r ] & \o_\delta . } \end{split}\ ] ] there is a nice result due to verdier , guaranteeing that a commuting square @xmath290\ar[d ] & y'\ar[d ] \\ x'\ar[r ] & y } \end{split}\ ] ] extends to a `` 9diagram '' of the form @xmath291\ar[d ] & y'\ar[r]\ar[d ] & z'\ar[r]\ar[d ] & x'\ar[d][1 ] \\ x\ar[r]\ar[d ] & y\ar[r]\ar[d ] & z\ar[r]\ar[d ] & x\ar[d][1 ] \\ x''\ar[r]\ar[d ] & y''\ar[r]\ar[d ] & z''\ar[r]\ar[d ] & x''\ar[d][1 ] \\ x'\ar[r][1 ] & y'\ar[r][1 ] & z'\ar[r][1 ] & x'[2]\ , , \\ } \end{split}\ ] ] where all the rows and columns are distinguished triangles , every square commutes , except for the last one ( containing the shift operator @xmath292 $ ] ) , which anticommutes ( for more details see page 24 in @xcite ) . applying verdier s `` 9diagram '' construction to ( [ c2 ] ) yields @xmath293\ar[d ] & { \cal s}_i \ar[r]\ar[d ] & { \cal c}_{i+1}\star { \cal s}_i \ar[r]\ar[d ] & { \cal t}_{i+1}\star { \cal s}_i [ 1]\ar[d ] \\ { \cal t}_{i+1 } \ar[r ] \ar[d ] & \o_\delta \ar[r ] \ar[d ] & { \cal c}_{i+1 } \ar[d ] \ar[r ] & { \cal t}_{i+1 } [ 1 ] \ar[d ] \\ { \cal t}_{i+1}\star ( { \cal c}_i \star \ldots \star { \cal c}_0 ) \ar[d ] \ar[r ] & { \cal c}_i \star \ldots \star { \cal c}_0 \ar[r ] \ar[d ] & { \cal c}_{i+1}\star ( { \cal c}_i \star \ldots \star { \cal c}_0 ) \ar[r ] \ar[d ] & { \cal t}_{i+1}\star ( { \cal c}_i \star \ldots \star { \cal c}_0)[1 ] \ar[d ] \\ { \cal t}_{i+1}\star { \cal s}_i[1 ] \ar[r ] & { \cal s}_i[1 ] \ar[r ] & { \cal c}_{i+1}\star { \cal s}_i [ 1 ] \ar[r ] & { \cal t}_{i+1}\star { \cal s}_i [ 2]\ , . \\ } \ ] ] we are interested in the term @xmath294 to compute it , return for a moment to the commutative diagram ( [ c34 ] ) , and consider the `` diagonal '' map @xmath295 the axioms of a triangulated category guarantee that the morphism @xmath296 can be included in a distinguished triangle of the form @xmath297.\ ] ] a crucial piece of the proof of verdier s `` 9diagram '' ( page 24 in @xcite ) provides another distinguished triangle that involves @xmath298 namely @xmath299 \to a[1]\,.\ ] ] returning to diagram ( [ c3 ] ) , we obtain the distinguished triangles @xmath300\ , , \ ] ] and @xmath301 \to { \cal x } [ 1]\ , , \ ] ] for some element @xmath302 in @xmath303 we plan on using the latter triangle to compute the object @xmath304 but for that we need to first understand the term @xmath305 the leftmost column of diagram ( [ c3 ] ) shows that @xmath306 by definition @xmath307 . after inspecting the definitions of the kernels @xmath308 and @xmath309 it is not hard to see that computing @xmath310 requires the calculation of kernels of the type @xmath311 with @xmath312 and @xmath313 the projection to a point . but @xmath314 since @xmath80 is ez spherical , the long exact cohomology sequence induced by the distinguished triangle ( [ eq : tr ] ) implies that @xmath315 summing up our work , we can conclude that @xmath316 where @xmath317 is the following complex in @xmath202 @xmath318 arranged such that the sheaf @xmath319 is located at the 0-th position . but _ what _ is the complex @xmath317 after all ? again the answer can be obtained by employing beilinson s resolution . consider the following two _ quasi - isomorphic _ complexes obtained by truncating ( [ eq : beil ] ) at the appropriate place : @xmath320 arranged such that the sheaf @xmath274 in the second complex is located at the 0-th position ( and , as a result , the sheaf @xmath321 is located at the @xmath322-th position in the first complex ) . call these complexes @xmath323 and @xmath324 respectively . on one hand , the well known properties of the cohomology groups of the projective space @xmath182 give that @xmath325 on the other hand , the same cohomology properties give that @xmath326.\ ] ] but @xmath327 and @xmath323 are quasi - isomorphic ( i.e. isomorphic in @xmath328 ) , hence @xmath329,\ ] ] and @xmath330\,.\ ] ] the distinguished triangle ( [ eq : triang2 ] ) and the definition ( [ eq : defss ] ) of the complexes @xmath267 show that in fact our unknown complex @xmath302 is nothing else but @xmath331 the proof by induction of the main claim of this section is then finished by invoking the distinguished triangle ( [ eq : triang1 ] ) . it was discussed above that the case of @xmath4 being a point amounts to a single d - brane becoming massless . this is the case originally studied by strominger in @xcite and yielding monodromy of the form studied in detail by seidel and thomas @xcite . there are three possibilities : 1 . @xmath103 in which case we are looking at the primary component of the discriminant . the quintic was studied at length in @xcite . 2 . @xmath3 is a complex surface of codimension one in @xmath2 . this could arise from the blow - up of an isolated quotient singularity . this was studied for example in @xcite . 3 . @xmath3 is a rational curve . this is the flop case and was studied in @xcite . now we have an infinite number of massless d - branes arising from the derived category of an algebraic curve . there are two possibilities : 1 . @xmath103 in which case @xmath2 is a k3-fibration and @xmath104 . this is the case we studied in section [ s : mon ] . @xmath3 is a ruled surface arising from blowing up a curve of quotient singularities in @xmath2 . in either case we are essentially looking at nonperturbatively enhanced gauge symmetry @xcite . putting @xmath332 for some point @xmath333 we obtain a soliton @xmath334 which is the structure sheaf of a single fibre @xmath335 of the map @xmath83 . these correspond to the charged vector bosons responsible for the enhanced gauge symmetry . clearly these bosons classically have a moduli space given by @xmath4 since @xmath336 may vary in @xmath4 . upon including quantum effects this leads to a number of massless hypermultiplets in the theory given by the genus of @xmath4 @xcite . putting @xmath337 we obtain @xmath338 which corresponds to one of the massless monopoles . in fact we may analyze the complete spectrum of solitons in seiberg witten theory @xcite by using the `` geometric engineering '' approach of @xcite to `` zoom in '' on the point in the moduli space where the nonabelian gauge symmetry appears . we will explain exactly what happens in detail in @xcite . it is worth speculating that such analysis of seiberg witten theory may shed some light on the `` local mirror symmetry '' story of papers such as @xcite . the massless b - type d - branes associated with the derived category of @xmath4 may be related to the a - type d - brane story of @xcite by some kind of local homological mirror symmetry . there is only one possibility , namely @xmath339 is an elliptic fibration over @xmath4 . we will denote this fibration @xmath340 . let @xmath341 correspond to the skyscraper sheaf of a point @xmath342 . then @xmath343 corresponds to the structure sheaf of an elliptic fibre @xmath344 over @xmath345 . let us consider the case where the size of @xmath108 becomes infinite . according to our rules then , the 2-brane wrapping @xmath346 should become massless when we hit the discriminant moving from the large radius phase to the phase where the elliptic fibres such as @xmath346 have collapsed . at first sight this looks peculiar . one does not usually expect a 2-brane wrapped around a 2-torus to become massless for a particular radius of the torus ! actually we will argue that this indeed happens and that it is when the 2-torus is zero sized that the 2-brane becomes massless . why t - duality does nt interfere with this will become apparent . as a specific example let @xmath2 be given by the following equation in @xmath347 : @xmath348 this has a quotient singularity which may be resolved with an exceptional divisor @xmath168 . one may regard @xmath349 $ ] as homogeneous coordinates on this @xmath350 . fixing a point on @xmath108 one then has an elliptic fibre @xmath346 given by a sextic in @xmath351 . see , for example , @xcite for more details about such fibrations . the moduli space of interest to us regards the area of @xmath346 . this area is infinite for the limit point and shrinks down as we approach the other limit point . for a more precise statement let us consider the mirror @xmath352 of @xmath346 given by the following equation in @xmath351 : @xmath353 varying the size of @xmath346 is then mirror to varying the complex structure of @xmath352 by varying @xmath354 . going through the usual story of solving the picard fuchs equation and using the mirror map to map back to the @xmath355 plane for @xmath346 we obtain the shaded region in figure [ f : wl ] . the result is that for the limit point we have @xmath356 and thus @xmath357 as expected . for the other limit point @xmath358 and @xmath359 . the discriminant is given by @xmath360 and corresponds to @xmath361 , i.e. , @xmath346 has zero area . this is where the 2-branes wrapping @xmath346 ( and all the other d - branes given by @xmath137 for @xmath341 ) become massless . so what about t - duality for @xmath346 ? note that the moduli space in figure [ f : wl ] corresponds to _ two _ fundamental regions for the action of @xmath362 on the upper - half plane . t - duality should map the point at @xmath360 to the point at @xmath363 . it is important to remember however that t - duality does not act only upon @xmath346 one must also shift the string dilaton by an amount related to the resulting change in the area of the torus . as such , the point at @xmath360 is related to @xmath363 only with the string coupling shifted off to infinity , implying again that the d - brane mass is zero . therefore if we fix the dilaton to be a finite value we can not use t - duality to relate @xmath360 to any large radius torus . this explains why we really do have a massless 2-brane appearing wrapped around a zero - sized torus . note also that if we allow the base @xmath350 to have finite size then the t - duality group ceases to exist anyway as in @xcite . the fact that there is a t - duality relating @xmath363 to @xmath360 shows that their physics must be similar . in particular @xmath360 must be an infinite distance away in the moduli space . indeed , in many respects the spectrum of stable d - branes at @xmath360 , coming from the derived category of @xmath168 , must be similar to the spectrum one would see upon going to a large radius limit . it would be interesting to investigate this in more detail and generality . finally let us note that not all the walls of the khler cone correspond directly to some subspace @xmath3 collapsing to @xmath4 . even so , it appears that we can still fit many , if not all examples into the general ez language . we illustrate this with an `` exoflop '' . let @xmath2 be the degree 12 hypersurface in @xmath364 . its mirror , @xmath75 , then has defining equation @xmath365 and we may use the following algebraic coordinates on the moduli space of complex structures of @xmath75 : @xmath366 we have chosen coordinates so that the khler cone of @xmath2 appears naturally as a positive octant in the secondary fan . in particular this means that the 3 rational curves in the moduli space connecting the large radius limit to each of the three neighbouring phase limit points are given by setting @xmath367 , @xmath368 or @xmath369 respectively . the discriminant has two components given by @xmath370 and @xmath371 the phase transition of interest occurs for the rational curve @xmath71 for which @xmath369 . for small @xmath101 we are in the phase . for large @xmath101 the @xmath2 undergoes an `` exoflop '' @xcite . that is @xmath2 becomes reducible with one component consisting of a threefold with a singularity . the other component consists of a fibration of a landau ginzburg orbifold theory over a @xmath96 . these components intersect at a point which is the singularity in the threefold component . this then is not an ez transformation . note however that the discriminant @xmath98 intersects @xmath71 transversely at @xmath372 and so the monodromy within @xmath71 is exactly given by monodromy around the primary component of the discriminant . therefore exactly one d - brane becomes massless at the transition point the d6-brane wrapping @xmath2 . this exoflop transition is equivalent , as far as monodromy is concerned , to an ez transformation with @xmath339 and @xmath4 given by a point . this is not entirely surprising given the following . classically one would describe the exoflop wall of the khler cone as a wall where @xmath373 . thus the classical volume of @xmath2 is going to zero , even though the volume of some surfaces and curves within @xmath2 , as measured by the khler form , do not vanish . since the volume of @xmath2 vanishes , one should expect the d6-brane to have vanishing mass . the fact that no other d - branes become massless is not obvious . it would be interesting to show that all phase transitions give rise to monodromies that can be associated with ez transformations . it is a pleasure to thank j. distler , d. morrison and r. plesser for useful conversations . is supported in part by nsf grant dms-0074072 and by a research fellowship from the alfred p. sloan foundation . r.l.k . was partly supported by nsf grants dms-9983320 and dms-0074072 .	we analyze the link between the occurrence of massless b - type d - branes for specific values of moduli and monodromy around such points in the moduli space . this allows us to propose a classification of all massless b - type d - branes at any point in the moduli space of s . this classification then justifies a previous conjecture due to horja for the general form of monodromy . our analysis is based on using monodromies around points in moduli space where a single d - brane becomes massless to generate monodromies around points where an infinite number become massless . we discuss the various possibilities within the classification . 0
the dynamics of the pacific ocean , within a band of about @xmath0 latitude from the equator , is characterised by a pronounced stratification , greater than anywhere else in the ocean ; see @xcite . a very sharp layer separates a shallow layer of relatively warm water ( and so less dense ) near the surface , from a deeper layer of denser , colder water . since these distinct top and bottom layers differ by density and temperature , the sharp layer between them is termed _ pycnocline _ and sometimes _ thermocline_. the difference in density across the thermocline is about 1% . in addition , underlying non - uniform currents occur , see e.g. @xcite . in a band approximately 300 km wide about the equator , the underlying currents are highly depth - dependent : in a sub - surface layer , typically extending no more than about 100 m down , there is a westward drift that is driven by the prevailing trade winds ; just below this there lies the equatorial undercurrent ( euc ) , an eastward jet whose core resides on the thermocline . below the euc , the motion dies out rapidly so that , at depths in excess of about 240 m , there is an abyssal layer of essentially still water . variety of waves in this region is observed , including long waves with wavelengths exceeding 100 km @xcite . there is ample evidence of large - amplitude internal waves , with relatively short wavelengths ( typically a few hundreds of metres ) and periods about 5 - 10 min ; see @xcite . we examine a simplified model of two - dimensional internal geophysical waves propagating along the equator . although the coriolis force plays the role of a waveguide keeping the euc and the occurring waves to propagate along the equator ( and to remain essentially two dimensional ) its impact on the dynamics of this two dimensional motion is rather small . thus , for simplicity we neglect coriolis forces . the gravitationally induced internal waves propagate on the common interface ( thermocline / pycnocline ) between the upper and lower media of different densities . ( we assume for simplicity that both media have constant densities . ) in physical reality the ocean surface will have surface waves . we neglect the surface waves and hence consider the top surface to be a flat boundary like a rigid lid . for long and intermediate surface waves coupled with internal waves such an approximation is justified , cf . @xcite . the boundary underneath is by an impermeable flat bed , which is a relevant approximation for the ocean bed , in the equatorial pacific . recently , explicit nonlinear solutions for equatorially trapped waves were obtained in the lagrangian framework , see @xcite . correspondingly , solutions for the internal waves have been described in @xcite . however , both these types of solution are restricted to relatively short wavelengths , and they are realistic representations of the flow only for the region near the surface and in the neighbourhood of the thermocline , respectively . by examining the governing equations of the system we provide a hamiltonian formulation of the equations and with a variable transformation we show that it has a canonical hamiltonian structure . the hamiltonian formulation of surface waves over deep water has been noticed by zakharov @xcite and since then the hamiltonian models have been used extensively for various approximations with linear and nonlinear pdes , e.g. @xcite . as per figure 1 we define the lower medium @xmath1 as the domain @xmath2 , the upper medium @xmath3 as the domain @xmath4 and the entire system @xmath5 as the domain @xmath6 where @xmath7 describes the elevation of the common interface . the subscript @xmath8 will be used to denote evaluation at the common interface . the shear flow in a general form is @xmath9 and moreover @xmath10 is a continuous function for @xmath11 and @xmath12 . it is apparent that @xmath13 . + the constants @xmath14 @xmath15 , @xmath16 , @xmath17 are positive ( see fig . 1 ) . we use the subscript notation @xmath18 to represent the lower and upper media respectively . the next assumption is that the pycnocline is always in the strip @xmath19 , i.e. @xmath20 for all @xmath21 so that all internal waves take place in this _ strip_. the velocities in the two media are decomposed into wave and current parts : @xmath22 where without any ambiguity @xmath23 the non - lateral velocity flow , with propagation in the positive @xmath24-direction , is given by @xmath25 @xmath26 and @xmath27 are the respective constant densities of the lower and upper media and stability is given by the condition that @xmath28 and @xmath3 with densities @xmath26 and @xmath27 . the shear profile @xmath29 is provided qualitatively . the five layers i v correspond to the description in @xcite . the important feature is that the wave motion takes place in layers ii and iii where the vorticity is zero and @xmath29 is constant . in the other layers @xmath29 is continuous . ] we assume that for large @xmath30 the amplitude of @xmath31 attenuates and hence make the following assumptions @xmath32 we write euler s equation as : @xmath33 where @xmath34 is the pressure at a depth @xmath35 , @xmath36 is ( constant ) atmospheric pressure , @xmath37 is the dynamic pressure due to the wave motion , @xmath38 is the acceleration due to gravity ( where @xmath35 points in the opposite direction to the center of gravity ) , @xmath39 is the force due to gravity per unit mass , and the coriolis force is neglected for simplicity . its effect is not significant and also it is important to keep the waves one - dimensional . applying equations ( [ phi_def ] ) and noticing that in the strip @xmath40 we have @xmath41 and velocities @xmath42 and thus there is no vorticity in the strip , i.e. @xmath43 . the equations can be written as @xmath44 where @xmath45 . at the interface @xmath46 therefore we have the conservation law ( bernoulli condition ) at the interface @xmath47 writing the bernoulli condition in terms of wave potentials @xmath48 only @xmath49 the difference with the case without current is only the shift @xmath50 this also suggests the definition of @xmath51 as the interface velocity potential and hence the definition of ] @xmath52 we will also use the following kinematic boundary conditions : @xmath53 noting that @xmath54 and @xmath55 , where the subscripts @xmath56 and @xmath57 denote evaluation at the bottom ( lower boundary ) and lid ( upper boundary ) respectively . again we observe only a shift @xmath58 in comparison with the case @xmath59 : @xmath60 therefore the time evolution of the variables , describing the dynamics of the internal waves @xmath61 does not depend on the shear outside the strip ; the constant flow in the strip leads only to a velocity shift with the velocity of the current @xmath15 ( in comparison to the currents free case ) . we will analyse the phenomenon from the hamiltonian point of view . if we consider the system under study as an irrotational system the hamiltonian , @xmath62 , is given by the sum of the kinetic and potential energies as : @xmath63 we try to express the hamiltonian in terms of @xmath64 . the kinetic energy of the lower layer is @xmath65 \ , dy\,dx \\= & \frac{\rho_1}{2}\ii \jj ( \nabla \varphi_1)^2 \ , dy\,dx + \rho_1 \ii \jj ( \varphi_{1,x } ) u_1(y ) \ , dy\,dx + \frac{\rho_1}{2}\ii \jj u_1 ^ 2(y ) \ , dy\,dx . \end{split}\ ] ] we evaluate @xmath66 since the integrals over total @xmath24 derivatives are zeroes due to the boundary conditions . the integral @xmath67 gives an irrelevant constant , which can be absorbed by a proper renormalization and does not produce any dynamic terms , since the average elevation of the thermocline by definition is zero at @xmath68 , i.e. @xmath69 the integrals over the upper layer produce @xmath70 and altogether the kinetic terms produce the extra term @xmath71 this quantity itself is a conservation law , related to the energy of the overall translation of the system by a velocity @xmath15 , see @xcite . the hamiltonian of the system is given by : @xmath72 we notice that the hamiltonian does not depend on the @xmath29 outside the _ strip_. in order to express it in terms of the desired variables we introduce the dirichlet - neumann operators @xmath73 given by @xmath74 where @xmath75 is the normal derivative of the velocity potential @xmath48 , at the surface , for an outward normal @xmath76 , see more details in @xcite . we define @xmath77 we can express the hamiltonian in terms of the conjugate variables @xmath78 as @xmath79 where @xmath80 is the hamiltonian in absence of shear , @xmath81 everywhere . from ( [ kbc ] ) , @xmath82 and can be written as a canonical hamiltonian system @xmath83 or @xmath84 for the shifted variables @xmath85 and @xmath86 , @xmath87 @xmath88 therefore this system of equations has the usual canonical hamiltonian form ( same as the hamiltonian form when there is no current ) after the galilean change of the coordinates , i.e. in a system travelling with velocity @xmath15 together with the flow in the strip @xmath89 . by taylor expanding the dirichlet - neumann operator we can represent it in terms of orders of @xmath31 as @xmath90 with the constant , linear and quadratic terms given as @xmath91 where the operator @xmath92 is a fourier multiplier equivalent to both the operation @xmath93 and the wavenumber @xmath94 , i.e. @xmath95 the operator @xmath96 can therefore be expressed as @xmath97 and , also , the hamiltonian can be represented as @xmath98 using explicit expressions @xmath99 in terms of @xmath31 and @xmath100 is given by @xmath101 recalling ( [ hsys ] ) we calculate the linearized equations of motion as @xmath102 looking for solutions in the form @xmath103 where , as usual @xmath94 is the wave number , @xmath104 is the angular frequency and @xmath105 is the wave speed , from the compatibility of the two equations we obtain the dispersion law @xmath106 in the case of infinite media as @xmath107 then @xmath108 and becomes @xmath109 the long - wave approximation of is the case @xmath110 which gives @xmath111 lastly , the consideration of the system as a single media system , i.e. @xmath112 in gives @xmath113 which is the well known dispersion relation for the linear approximation of gravity water waves in a single medium irrotational system @xcite . the shear flow in the layers , away from the thermocline ( the strip where the internal waves occur ) does not affect the dynamics of the internal waves . the results are complementary to the case where the vorticity is constant but with a finite jump at @xmath114 considered in @xcite . in the latter case vorticity does affect the wavespeed of the internal waves , cf . with @xcite . a.c . is funded by the fiosraigh scholarship programme of dublin institute of technology . the support of the fwf project i544-n13 `` lagrangian kinematics of water waves '' of the austrian science fund is gratefully acknowledged by r.i . the authors are grateful to prof . a. constantin for many valuable discussions and to an anonymous referee for several very useful suggestions . a. compelli , hamiltonian approach to the modeling of internal geophysical waves with vorticity , _ monatsh . math . _ ( 2014 ) , published online , doi:10.1007/s00605 - 014 - 0724 - 1 ; available from http://arrow.dit.ie / scschmatart/178/. v. zakharov , stability of periodic waves of finite amplitude on the surface of a deep fluid , _ zh . fiz . _ * 9 * ( 1968 ) 8694 ( in russian ) ; _ j. appl . mech . tech . phys . _ * 9 * ( 1968 ) 190194 ( english translation ) .	the interaction of the nonlinear internal waves with a nonuniform current with a specific form , characteristic for the equatorial undercurrent , is studied . the current has no vorticity in the layer , where the internal wave motion takes place . we show that the nonzero vorticity that might be occuring in other layers of the current does not affect the wave motion . the equations of motion are formulated as a hamiltonian system . _ keywords _ : internal waves , equatorial undercurrent , shear flow , hamiltonian system . _ 2000 mathematics subject classification _ : 35q35 , 37k05 , 74j30
recently , gauge / gravity duality @xcite has been used to model strongly - interacting systems in terms of a gravity dual . the most important case is the strongly - interacting quark gluon plasma created at rhic . while a systematic treatment is still lacking , the @xmath1 super - yang - mills plasma has been used instead , with some success . most notably , the viscosity / entropy density ratio , universal among all theories with a gravity dual @xcite , seems to be not too far away from the value extracted from experimental data @xcite . the long - distance behavior of a black hole horizon is captured by the same hydrodynamic equation governing the evolution of the plasma . as a result , from the gravity equations one can reconstruct the hydrodynamic equations , including transport coefficients . one can go beyond leading order in the gravity equations and capture second - order corrections , in the process discovering extra terms typically ignored in almost all implementations of the israel - stewart formalism @xcite . the approach has been extended to r - charged black holes , dual to fluids with chemical potentials @xcite . in this work , we use black holes to study relativistic superfluids . our work is a continuation of ref . @xcite , where a holographic model of a superfluid was constructed . here , we investigate the behavior of the superfluid when a superfluid current flows through the system . it is known that in nonrelativistic superfluids there is a critical superfluid velocity , above which the superfluid phase does not exist . we discover exactly the same phenomenon in the relativistic superfluids at low temperatures : there is a first - order phase transition between the superfluid and the normal phase as one changes the superfluid current . the phase transition becomes second order at higher temperatures . we will also be discussing the superfluid hydrodynamics of such systems , in particular , the speed of second sound . a short discussion of hydrodynamics can be found in @xcite . hydrodynamics of a system with spontaneously broken symmetries is different from hydrodynamics of normal liquids because the system contains long - range modes ( goldstone bosons ) which must be included in the hydrodynamic equations . let us first review the thermodynamics of quantum systems with a spontaneous @xmath0 symmetry breaking . for simplicity consider a system with one global @xmath0 symmetry , such as complex @xmath2 theory . we define the free energy @xmath3 where z = tr e^-h , and @xmath4 is the hamiltonian . while we eventually take the thermodynamic limit @xmath5 , for now it is more convenient to think of the system as living inside a @xmath6 dimensional torus @xmath7 where the perimeters of the @xmath6 circles are @xmath8 . when the @xmath0 symmetry is not spontaneously broken , the free energy @xmath9 depends only on the temperature , and , if we do not take the thermodynamic limit , the @xmath8 : @xmath10 . when the symmetry is spontaneously broken , the system contains a massless ( pseudo ) scalar field @xmath11 , which is compact : @xmath12 . in the example of the complex @xmath2 theory , @xmath11 is just the phase of @xmath13 . if one thinks about @xmath11 as the elementary field , periodicity in @xmath11 space means that we are free to impose boundary conditions such that @xmath14 . thus there is a set of partition functions @xmath15 $ ] , where we integrate over all fields with boundary conditions specified by @xmath16 . the usual ground state corresponds to @xmath17 . as we shall see later , @xmath15 $ ] does not exist in the strict thermodynamic sense . physically , if one sets @xmath18 , the system will relax over time to a state with no winding . however , the time scale for this process may be long , and the winding state is practically stable . we assume stability . define @xmath19 , an analogue of the topological charge density whose integral can distinguish various boundary conditions : @xmath20 where @xmath21 is the @xmath22th spatial circle of the @xmath7 . we can therefore impose the boundary conditions on @xmath11 by introducing lagrange multipliers @xmath23 which pick out the required values of @xmath24 : @xmath25 = \int_{\{\n\}}\!\ ! d\phi \ ; e^{-s_e[\phi ] } = \int\!\ ! d\phi\ , d\lambda \ ; e^{-s_e[\phi]}\ ; \exp \biggl ( i \sum_j \lambda_j \biggl ( \oint_{c_j } \q \cdot \dx - 2\pi n_j \biggr ) \biggr ) \ .\ ] ] we are taking advantage of the fact that the euclidean path integral on a compact time circle of perimeter @xmath26 yields the thermal partition function at a temperature @xmath27 . having compactified the time direction , it is natural to ask what boundary conditions should be applied for @xmath11 in the time like direction . the answer to this question involves introducing a chemical potential , and we will return to it shortly . we can write the field @xmath11 as @xmath28 where @xmath29 is constant , and @xmath30 is periodic . the delta function in the path integral for @xmath15 $ ] will pick out @xmath31 , so knowing @xmath29 is equivalent to knowing the boundary conditions for @xmath11 . in the thermodynamic limit , @xmath9 depends on the @xmath8 only through an overall multiplicative volume factor @xmath32 and @xmath29 becomes a continuous variable which transforms as a vector under rotations . at zero temperature , the energy ( for given @xmath16 ) is always minimized for @xmath33 because that s when the @xmath34 term is smallest . so at zero temperature , the ground state can be uniquely characterized by the @xmath29 . we assume now that the thermal state can be uniquely characterized by @xmath29 and @xmath27 , and define the partition function as @xmath35 = \int\!\ ! d\phi\ , d\lambda \ ; e^{-s_e[\phi]}\ ; \exp\biggl ( i \sum_j \lambda_j \oint_{c_j } ( \grad \varphi - \al ) \cdot \dx \biggr ) \ , . \label{eq : z}\ ] ] note that @xmath36 is the equilbrium value of @xmath37 . in nonrelativistic superfluids @xmath38 is proportional to the superfluid velocity . we shall use the same terminology and call @xmath38 the superfluid velocity . @xmath39 $ ] describes a stationary state with constant , non - zero superfluid velocity . the partition function @xmath40 is a scalar , and therefore can only depend on @xmath41 in the thermodynamic limit . note that the state described by @xmath39 $ ] is in fact metastable , rather than stable . this is because @xmath11 , as the phase of a complex field @xmath13 , is ill - defined when @xmath42 ; hence the winding numbers @xmath24 are not topological and can change with time . the process of `` unwinding '' can be visualized as follows . imagine a state with only one of the @xmath24 equal to 1 , and other winding numbers equal 0 . the state can be thought of as containing a @xmath43 dimensional domain wall , across which @xmath11 changes by @xmath44 . however , in @xmath45 dimensions , there is no topological conservation law which would ensure the stability of the domain wall , and the domain wall will decay through hole nucleation . in terms of superfluid hydrodynamics in 3 + 1 dimensions , nucleating a hole in the two - dimensional domain wall corresponds to producing a vortex loop . production of vortex loops violates landau s criterion for superfluidity @xcite , and renders the state with non - zero @xmath36 metastable . in a theory with large @xmath46 , like the one we will be considering , the rate of vortex loop production is exponentially suppressed , as both the energy of a vortex loop , and the action of the configuration describing vortex nucleations , are proportional to @xmath46 . therefore , provided that the volume and time scales are not too large , one can treat states with a nonzero superfluid velocity as thermal equilibrium states . we now would like to introduce a chemical potential @xmath47 and work in the grand canonical ensemble . we define the potential function @xmath48 where = tr e^-(h - q ) [ eq : calz ] and @xmath49 is the conserved charge corresponding to the @xmath0 symmetry . to introduce @xmath47 at the level of the path integral , it is useful to gauge the @xmath0 symmetry by coupling the system to a non - dynamical gauge field @xmath50 : [ a ] = e^w[a ] = d e^-s_e [ , a ] . [ eq : za ] we are temporarily ignoring the dependence of @xmath51 on the @xmath36 . the one point function of the u(1 ) current is then generated by @xmath52 $ ] : = j^(x ) . [ onepoint ] in a system where the external field strength is zero @xmath53 , we may pick a gauge in which @xmath50 is constant and it makes sense , given ( [ onepoint ] ) , to interpret @xmath54 as the chemical potential . gauge transformations which send @xmath55 have a nontrivial effect on the compact scalar : @xmath56 . in the presence of the background gauge field , the pressure @xmath57 can depend only on gauge invariant quantities such as @xmath58 . the grand canonical partition function thus becomes [ a , ] = d d e^-s_e [ , a ] ( i _ j _ j _ c_j ( -a- ) ) . [ eq : calzpartition ] like @xmath39 $ ] , by rotational symmetry @xmath59 $ ] can only depend on @xmath41 in the thermodynamic limit . we are assuming that the pressure @xmath60 depends on three parameters , @xmath61 . we can now choose a gauge in which @xmath62 , so that @xmath63 . the chemical potential @xmath47 can be interpreted as the zero component of the gauge field , and we have @xmath64 . at zero temperature , the system is lorentz invariant , provided one treats @xmath65 and @xmath66 as spurion fields that transform as a four - vector under the lorentz group . as a result , the pressure can only depend on @xmath67 . at finite temperature , the pressure depends on @xmath65 and @xmath68 separately , which we can write as @xmath69 . in other words , gauging the symmetry allows us to trade the dependence of pressure on the condensate phase for the dependence of pressure on the background gauge field . in a general gauge , @xmath70 . denoting @xmath71 , we have the pressure as a function of three thermodynamic variables : @xmath72 . from it one can define the conjugate variables , @xmath73 notice that @xmath74 , @xmath75 , and @xmath76 are functions of @xmath27 , @xmath47 , and @xmath77 . if the third term were absent from the eq . ( [ eq : dp ] ) , then @xmath74 would be the entropy density , and @xmath75 the charge density . we will give the interpretation of @xmath74 , @xmath75 and @xmath76 in the next section . let us set the external gauge field @xmath50 to 0 . the equations of ideal relativistic superfluid hydrodynamics are known . we will use the form written in @xcite . in this formulation , the degrees of freedom are @xmath27 , @xmath47 , @xmath11 , and a unit four - vector @xmath78 satisfying @xmath79 ( recall that we use the mostly plus convention for the flat metric tensor @xmath80 ) . we will later identify @xmath78 with the velocity of the normal component . the equations consist of the conservations of the energy - momentum tensor @xmath81 and of the @xmath0 symmetry current @xmath82 , @xmath83 and a `` josephson equation '' describing time evolution of @xmath11 , @xmath84 the stress - energy tensor and the current are expressed in terms of the hydrodynamic variables through @xmath85 where the energy density @xmath86 is defined by @xmath87 . thus we have @xmath88 hydrodynamic equations ( [ eq : tmunu - conservation ] ) , ( [ eq : jmu - conservation ] ) , ( [ eq : mu - definition ] ) for @xmath88 variables @xmath27 , @xmath47 , @xmath11 and @xmath78 . one can derive from eqs . ( [ eq : tmunu - conservation])([eq : jmu - definition ] ) @xmath89 one can interpret this equation as the equation of entropy conservation ( entropy is conserved since we are working at the level of ideal , nonviscous hydrodynamics ) . thus @xmath74 is the entropy density and @xmath78 is the velocity of entropy flow . in the two - fluid model only the normal component carries entropy ; therefore @xmath78 is interpreted as the normal velocity . the two contributions to the current in eq . ( [ eq : jmu - definition ] ) can be interpreted as the normal and superfluid currents . therefore @xmath75 is the normal density , and @xmath76 is the analogue of the pion decay constant . ( the equivalent of the superfluid density would be @xmath90 . ) let us look at small fluctuations about an equilibrium state at fixed temperature , chemical potential and zero normal and superfluid velocities , i.e. we write @xmath91 , @xmath92 , @xmath93 , @xmath94 , where @xmath95 , @xmath96 , @xmath97 , and @xmath98 are small . in terms of @xmath99 the variation of pressure is @xmath100 , where @xmath101 . further , let us assume that pressure is a smooth function of @xmath102 at small @xmath99 , so that @xmath103 in equilibrium . the linearized hydrodynamic equations become ( omiting subscript `` 0 '' on equilibrium quantities ) [ eq : linear - sf - hydro ] @xmath104 in this system of equations , pressure is taken as @xmath105 , and @xmath106 is the density of enthalpy . the first equation is the linearized current conservation equation @xmath107 . the second equation is the linearized energy conservation @xmath108 . the third equation is the linearized momentum conservation , @xmath109 . finally , the fourth equation says that @xmath47 and @xmath110 are not independent because @xmath111 while @xmath112 . an interesting feature of the linearized hydrodynamic equations ( [ eq : linear - sf - hydro ] ) is that they admit propagating mode solutions even if one ignores the fluctuations of the energy - momentum tensor , i.e. if one ignores the ( normal ) sound fluctuations . if one were to ignore the condition @xmath113 , together with fluctuations of temperature @xmath95 and velocity of the normal component @xmath114 , then the system ( [ eq : linear - sf - hydro ] ) becomes @xmath115 [ eq : linearprobe ] fourier transforming all variables as @xmath116 , we find a propagating mode with frequency @xmath117 when expressing @xmath76 in terms of @xmath118 , we have assumed that @xmath119 at small @xmath99 . the thermodynamic derivatives in eq . ( [ eq : second - sound - leftover ] ) are to be evaluated at @xmath120 . the propagating mode ( [ eq : second - sound - leftover ] ) is the leftover of the second sound in superfluids which survives even if one ignores the @xmath113 part of the hydrodynamic equations . one expects that this mode is captured by the dual gravitational description which ignores the backreaction of the gauge fields on the metric . it is not difficult to find the eigenmodes of the full system ( [ eq : linear - sf - hydro ] ) . taking all variables proportional to @xmath116 , one finds a fourth order equation for frequency , @xmath121 where the coefficients @xmath122 , @xmath123 and @xmath124 are independent of @xmath125 . when @xmath126 , the equation for @xmath127 has two real positive roots @xmath128 and @xmath129 . the first solution is the normal sound , and the second solution is the second sound . in terms of thermodynamic derivatives , the coefficients @xmath122 , @xmath123 and @xmath124 are given by @xmath130\,,\\ & & b = \left(\frac{\d^2 p}{\d t^2}\right)\ ! t ( n^2{+}wf^2 ) + \left(\frac{\d^2 p}{\d \mu^2}\right ) \ ! t s^2 - 2\left(\frac{\d^2 p}{\d t\d\mu}\right)\ ! ts\,n \ , , \\ & & c = f^2 t s^2\,.\end{aligned}\ ] ] in particular , @xmath122 , @xmath123 , and @xmath124 are positive . let us now look at simple examples . in the non - superfluid phase @xmath131 , and therefore @xmath132 , while the second sound is absent , @xmath133 . even in the non - superfluid phase , the speed of the normal sound looks like a complicated expression in terms of the derivatives of pressure @xmath134 . the expression for @xmath135 simplifies if instead of @xmath134 we work with @xmath136 indeed , the coefficient @xmath122 is proportional to the jacobian of the transformation from the @xmath137 variables to the @xmath138 variables . in terms of @xmath136 , the speed of sound in the normal phase becomes @xmath139 now that we have @xmath135 expressed in terms of @xmath136 without reference to the chemical potential , we can evaluate the thermodynamic derivatives in the canonical ensemble instead of the grand canonical . in the canonical ensemble , the total charge ( number of particles ) @xmath140 is fixed , and therefore @xmath141 . it will be convenient to go from the @xmath137 to the @xmath142 variables . for the total entropy @xmath143 we have @xmath144 while on the other hand the relation @xmath145 gives @xmath146 this implies that the speed of sound in the normal phase ( [ eq : vs - sn ] ) can be written as @xmath147 as another example , consider a scale - invariant theory , in which case pressure has the form @xmath148 , where @xmath149 is a dimensionless function . the speed of normal sound evaluated from eq . ( [ eq : w4 ] ) is @xmath150 , in either normal or superfluid phase . the speed of the second sound evaluated from eq . ( [ eq : w4 ] ) can be written as @xmath151^{-1}\,.\ ] ] we can see that in scale invariant theories , the cartoon expression ( [ eq : second - sound - leftover ] ) can be recovered in the formal `` large - entropy '' limit @xmath152 , together with @xmath153 . in theories which do not have scale invariance , the speed of the second sound does not have the simple form ( [ eq : second - sound - cft ] ) , but can be easily determined from eq . ( [ eq : w4 ] ) given the equation of state @xmath134 . note that the simple expression ( [ eq : second - sound - leftover ] ) ceases to be a good approximation to the speed of the second sound at @xmath154 . this is because at small temperatures we expect that the system can be described as a gas of massless goldstone bosons , with entropy density @xmath155 . on the other hand , we expect in the low - temperature region that @xmath156 which imples that the condition of large entropy can not be satisfied at @xmath154 . at low temperatures ( high densities ) the speed of the second sound has to be determined from the full equation ( [ eq : w4 ] ) . on the gravity side , we study the einstein - maxwell theory with a complex scalar field . fluctuations of the bulk metric correspond to fluctuations of @xmath81 on the boundary , while fluctuations of the @xmath0 gauge field correspond to fluctuations of @xmath157 on the boundary . the scalar field is charged under the bulk @xmath0 , and its background value corresponds to the condensate on the boundary . the mass of the scalar is a free parameter . we also need to specify the boundary conditions . the metric will be asymptotically ads because we want to study the superfluid system using the ads / cft correspondence . the boundary conditions for the scalar correspond to the `` normalizable '' mode because we want to describe the system in which a charged operator has a vev . however , the boundary conditions for the @xmath0 gauge field should correspond to the `` non - normalizable '' mode because we want to describe the system in the background gauge field , as discussed above . in other words , we fix the value of @xmath50 at the asymptotic ads infinity . fixing the value of @xmath65 at infinity amounts to fixing the chemical potential in the boundary theory , and leads to a charged black hole in ads . fixing the value of @xmath66 at infinity does not introduce additional `` hair '' for the black hole because we expect the new black hole solution to be only metastable , in accord with field theory expectations . so we want to find a ( metastable ) stationary solution with the above boundary conditions , and then study small fluctuations around this solution , corresponding to hydrodynamic fluctuations in the boundary theory . in the following , we will ignore the backreaction of the gauge and scalar fields on the metric . it would be nice to include the backreaction , at least perturbatively to leading order . the action is @xmath158\ , , \label{gravaction}\ ] ] where @xmath159 , @xmath160 , and capital latin indices run from @xmath161 to @xmath162 . the equations of motion are @xmath163 @xmath164 where the current is @xmath165 $ ] . we write the bulk scalar as @xmath166 , and make a gauge transformation @xmath167 . in the new gauge , the phase @xmath11 disappears from the equations of motion , and the current becomes @xmath168 from the maxwell equations ( [ eq : a - eqn ] ) one can see that the background value for @xmath169 would induce a ( position - dependent ) mass for the gauge field . this is the higgs mechanism in the bulk . we take the @xmath170 dimensional background metric to be of the following form : @xmath171 the metric with @xmath172 corresponds to pure ads and @xmath173 corresponds to our black hole solution . to find the background solution for @xmath169 and @xmath174 , we take all fields independent of @xmath175 and @xmath176 . the @xmath177-component of the maxwell equations ( [ eq : a - eqn ] ) now gives @xmath178 which means we can choose @xmath179 . the equations of motion become @xmath180 = \left(a_i^2 - \frac{a_t^2}{f } + \frac{m^2}{z^2 } \right)\rho\,,\ ] ] @xmath181 = \frac{e^2}{z^2 f}\ , \rho^2 a_t\ , , \label{ateq}\ ] ] @xmath182 = \frac{e^2}{z^2}\ , \rho^2 a_i\ , . \label{aieq}\ ] ] in the limit when @xmath183 they reduce to the coupled ( @xmath184 , @xmath169 ) system of equations studied recently in @xcite . note that we have a coupled system of non - linear odes it would be interesting to see if the solutions may exhibit chaotic behavior . up to boundary counter terms , the free energy of the field theory is determined by the value of the action ( [ gravaction ] ) evaluated on - shell , @xmath185 , where the ellipsis denotes boundary terms that we will presently introduce . employing the equations of motion , we may rewrite ( [ gravaction ] ) as s_os = d^dx . we have included a cut - off @xmath86 because this on - shell action is naively divergent and needs to be regularized through the addition of boundary terms . it is difficult to treat the general case both succinctly and clearly and several full treatments already exist in the literature ( see for example @xcite ) . we proceed to regularize this action in the simple case @xmath186 and @xmath187 . using the explicit form of the metric ( [ metric ] ) , the on - shell action reduces to s_os = d^d x . the near boundary behavior of the fields takes the form @xmath188 these boundary values have various reinterpretations in the field theory . for the gauge field , @xmath189 is the chemical potential while @xmath190 is a superfluid velocity . then @xmath191 is proportional to the charge density while @xmath192 are charge currents . for the scalar field , there exists an ambiguity @xcite . for a scalar operator @xmath193 of conformal dimension one , @xmath194 while @xmath123 is a source . for a scalar @xmath195 of conformal dimension two @xmath196 while @xmath122 is a source . in regulating @xmath197 , we must carefully formulate the boundary conditions . for example , do we wish to keep @xmath50 or @xmath198 fixed on the boundary ? keeping @xmath65 fixed corresponds to keeping the chemical potential fixed and thus working in the grand canonical ensemble in the field theory . in the spatially homogenous case where we can set @xmath179 , varying @xmath199 in the bulk leads to a boundary term proportional to @xmath200 . without an additional boundary term , we are working in an ensemble where @xmath201 is held fixed . if we would like to work in the canonical ensemble , at fixed charge , it is @xmath202 that must be held fixed at the boundary ( @xmath203 ) . to accommodate this change , we would need to make what amounts to a legendre transform and add the boundary term @xmath204 to the action ( [ gravaction ] ) . we will work in an ensemble where @xmath201 is held fixed and thus need no such further boundary terms . a similar decision needs to be made about the scalar operator . it is most natural to work in an ensemble where the value of @xmath205 is fixed instead of the source for the operator . however , we still must decide whether we want our scalar operator to have conformal dimension one or two . as the ensemble where the source for @xmath193 is fixed is equivalent to the ensemble where @xmath206 is fixed , we know that the ensembles where @xmath207 and @xmath206 are fixed must be related by a legendre transform @xcite . consider first the case where @xmath207 is fixed . the on - shell action is naively divergent , and we must add a counter - term . the counter - term and the @xmath208 term in the on - shell action combine to give @xmath209 in the legendre transformed case , where @xmath206 is fixed , we add two counter - terms , one to ensure that we hold @xmath210 fixed at the boundary instead of @xmath211 and one to control the divergence : @xmath212 recalling that this regularized on - shell action is the negative of the free energy and assuming a spatially homogenous system so that we may divide out by a factor of the volume @xmath32 , we find for the free energies that _ i ( , , o_i)/v = . [ omega ] recall @xmath213 . having fixed @xmath214 , @xmath47 and @xmath99 , the conjugate boundary values @xmath215 , @xmath49 , and @xmath216 are then determined through the dynamics of the gravitational theory . since @xmath217 and @xmath218 are legendre transforms of each other , we have @xmath219 thus critical points of the free energies correspond to gravitational solutions where at least one of the two @xmath214 vanish . \a ) b ) the nonlinear differential equations ( [ rhoeq][aieq ] ) appear to be intractable analytically . however , it is relatively straightforward to integrate the equations numerically . the results of this section are most succinctly summarized by the two phase diagrams for the scalars @xmath193 and @xmath195 shown in figure [ fig : pd ] . in this section . ] to see where these phase diagrams come from , we begin by reviewing the case @xmath120 studied in the canonical ensemble in @xcite while here we choose to work at fixed @xmath47 . given @xmath120 , the third differential equation ( [ aieq ] ) drops out . plots of the expectation value @xmath205 versus temperature are shown in figure [ fig : condensate ] , as the black curves on the far right . at high temperature , @xmath220 , but at the critical temperature @xmath221 , there is a second order phase transition where the expectation values become nonzero . for @xmath193 , @xmath222 while for @xmath195 , @xmath223 . near but slightly below @xmath221 the scalars exhibit the standard mean field scaling with the reduced temperature @xmath224 the most straightforward way to see that the phase transition is second order is to examine a plot of the free energy versus temperature : @xmath225 is smooth at @xmath221 . using eq . ( [ omega ] ) , we have produced figure [ fig : phasetransition]a for the scalar @xmath193 , which indeed shows this smooth behavior . we do not show a very similar plot for the second scalar @xmath195 . \a ) b ) \a ) b ) a more elaborate demonstration that the phase transition is second order comes from an investigation of @xmath225 as a function of the order parameter @xmath214 near @xmath221 , recovering completely standard results in the landau - ginzburg mean field theory of phase transitions . we have found numerically that = _ i(t ) ( ) ^2 + _ i(t ) ( ) ^4 fits the free energy curves extremely well . moreover , @xmath29 and @xmath226 are nearly linear in @xmath27 : @xmath227 by definition , @xmath29 passes through zero at the phase transition . as we increase the superfluid velocity @xmath99 , nothing dramatic happens immediately . the phase transition remains second order although @xmath221 decreases as can be seen from figure [ fig : condensate ] . because the decrease is due to the additional kinetic energy of the system , we expect the decrease to be quadratic in @xmath99 , which is born out by the shape of the second order lines in the phase diagram , figure [ fig : pd ] . numerically , we find that @xmath228 where @xmath229 for @xmath193 and @xmath230 for @xmath195 . however , there exists a critical @xmath99 above which the phase transition becomes first order . for @xmath193 , this critical velocity is @xmath231 while for @xmath195 it is @xmath232 . from figure [ fig : condensate ] , it is clear that something interesting must happen because the curves @xmath205 become multi - valued for sufficiently large @xmath99 . it is possible to see that the phase transition is first order in different ways . the simplest is to look at the free energy as a function of temperature . figure [ fig : phasetransition]b presents this classic swallow tail shape for @xmath193 and @xmath233 . at the phase transition , the free energy is continuous but not differentiable . the two nonanalytic points in the free energy curve are spinodal points " or points beyond which one of the phases ceases to exist even as a metastable minimum of the free energy . to demonstrate more convincingly that the phase transition becomes first order , we computed the free energy as a function of the order parameter near the putative tricritical point . we found numerically that the free energy is well described by the sixth order polynomial = _ i(t , ) ( ) ^2 + _ i(t , ) ( ) ^4 + _ i(t , ) ( ) ^6 . at the tricritical point , @xmath29 and @xmath226 both vanish . moreover , near the tricritical point , @xmath29 and @xmath226 vary linearly with @xmath27 and @xmath99 while @xmath234 is positive and roughly constant . \a ) , computed by evaluating thermodynamic derivatives in eq . ( [ eq : second - sound - leftover ] ) : a ) @xmath193 scalar , b ) @xmath195 scalar . the speed of second sound vanishes as @xmath235 and appears to approach a constant value as @xmath236.,title="fig:",width=254 ] 0.2 in b ) , computed by evaluating thermodynamic derivatives in eq . ( [ eq : second - sound - leftover ] ) : a ) @xmath193 scalar , b ) @xmath195 scalar . the speed of second sound vanishes as @xmath235 and appears to approach a constant value as @xmath236.,title="fig:",width=263 ] before moving to a discussion of second sound , we would like to point out one nice feature of fig . [ fig : condensate ] : the curves approach each other at low temperature . this agreement is a consequence of the fact that at @xmath237 , the pressure @xmath57 can only be a function of the lorentz invariant quantity @xmath238 , as discussed in section [ sec : thermo ] . the speed of second sound as a function of temperature is shown in fig . [ fig : v2-vs - t ] . the plots were computed in the probe approximation using eq . ( [ eq : second - sound - leftover ] ) . we have set the superfluid velocity @xmath120 . the behavior close to @xmath221 is qualitatively similar to that of superfluid @xmath239he @xcite . as the temperature is decreased from @xmath221 , the speed of second sound rises rapidly from zero and eventually levels off . experimentally , it is difficult to go to very low temperature and remain within the hydrodynamic limit . the scattering length for the phonons approaches the system size . numerically , we have also had difficulty finding solutions at very low temperatures ; the curves in fig . [ fig : v2-vs - t ] terminate where our numerics fail . theoretically , however , the expectation for @xmath239he is that second sound at @xmath240 is a sound wave supported by a gas of phonons @xcite . thus , the speed of second sound close to @xmath237 should approach @xmath241 . the naive extrapolation of the curves in fig . [ fig : v2-vs - t ] suggest that @xmath242 is @xmath243 in the @xmath193 case and @xmath244 in the @xmath195 case . the fact that the curves in fig . [ fig : v2-vs - t ] level off at low temperature is caused numerically by a similar growth in the susceptibility and the pion decay constant in eq . ( [ eq : second - sound - leftover ] ) . that we work in the probe approximation suggests our model may not be reliable at low temperature anyway . we have neglected the coupling between our abelian - higgs sector and the metric . the abelian - higgs sector on its own does not support ordinary sound . we have discussed the hydrodynamics of relativistic superfluids . moreover , we presented a holographic model that reproduces many of the familiar features of superfluids , including a critical superfluid velocity above which the system returns to its normal phase and a second order phase transition at zero superfluid velocity between the normal and superfluid phases . in many respects , our holographic model is similar to the landau - ginzburg mean field treatment of phase transitions . near the phase transition , we saw that the potential function for the order parameter has the familiar polynomial expansion @xmath245 . in our case , however , the values of @xmath246 and @xmath247 and indeed the entire structure of @xmath248 are not set directly by the field theorist but are instead encoded in a nontrivial fashion by the bulk gravitational solution . by assuming a simple gravitational model , we are led to a particular @xmath248 that is valid not just near @xmath221 but at all temperatures in the superfluid phase . moreover , from the model one can extract not only the static thermodynamic properties , but also quantities relevant for time - dependent , dynamic processes , like the kinetic coefficients and the correlation functions . in the future , it would be very interesting to extend the numerical results above . two obvious directions present themselves . in the calculation of the speed of second sound from derivatives of the pressure @xmath57 , one could go beyond the probe approximation and include the back reaction of the abelian - higgs sector on the metric . we hope that the resulting speed of second sound will be phenomenologically meaningful not only near @xmath249 but all the way down to @xmath237 . another interesting project for the future would be to find the sound wave poles in the density - density correlation function . by studying fluctuations of the scalar and gauge field of the form @xmath250 in the probe limit , one should be able to isolate a pole of the form @xmath251 in the fourier transform of the retarded green s function for the charge density . here the ellipsis denotes higher order ( damping ) terms in @xmath125 . beyond the probe limit , there should also be a pole corresponding to the propagation of ordinary sound . 0.2 in _ note added _ while we were completing this paper , we learned of ref . @xcite which has some overlap with this work . 0.2 in _ acknowledgments _ we would like to thank david huse for discussion . p.k.k . and c.p.h . thank the int for hospitality during the 2008 workshop from strings to things , " where this work began . thanks the galileo galilei institute for theoretical physics for hospitality and the infn for partial support during the completion of this work . the work of c.p.h . , p.k.k . , and d.t.s . was supported , in part , by the us nsf under grant no . phy-0756966 , nserc of canada , and the us doe under grant no . de - fg02 - 00er41132 , respectively . in this appendix we show that the hydrodynamic equations written in sec . [ sec : hydro ] are equivalent to the set of equations proposed previously by israel , carter , khalatnikov , and lebedev @xcite . the formulations of ref . @xcite and ref . @xcite have been shown to be equivalent in refs . we will follow the notation of ref . @xcite . in ref . @xcite , superfluid hydrodynamics is formulated as follows . first , one postulates that the thermodynamic properties of the superfluid are defined by a scalar function @xmath252 , which is a function of two 4-vectors @xmath253 and @xmath254 , which are the entropy density and the particle number density . since @xmath252 is a lorentz scalar , the number of variables that it depends on is three : @xmath255 , @xmath256 , and @xmath257 . at the first sight , eqs . ( [ ckeq1])([ckeq4 ] ) do not bear much resemblance to eqs.([eq : tmunu - conservation])([eq : s - conservation ] ) . however , these systems of equation are in fact equivalent . the variables appearing in the carter - khalatnikov formulation can be identified with the hydrodynamic variables used in this paper as in the following relations , @xmath262 \ . \label{cktheta}\end{aligned}\ ] ] we now show that any solution to the hydrodynamic equations ( [ eq : tmunu - conservation])([eq : s - conservation ] ) satisfies eqs . ( [ ckdlambda])([ckeq4 ] ) , upon the subsitutions ( [ cklambda])([cktheta ] ) . first , eqs . ( [ ckeq1 ] ) and ( [ ckeq2 ] ) coincide with eqs . ( [ eq : s - conservation ] ) and ( [ eq : jmu - conservation ] ) due to eqs . ( [ cks ] ) and ( [ ckn ] ) . furthermore , eq . ( [ ckeq3 ] ) is trivially satisfied by eq.([ckmu ] ) . now let us check eq . ( [ ckdlambda ] ) . the left hand side is @xmath263 where the thermodynamic relation @xmath264 and eq . ( [ eq : dp ] ) have been used , while the right hand side is @xmath265 ( u^\mu ds + sdu^\mu ) - \d_\mu\varphi \bigl [ u^\mu dn + ndu^\mu + d\bigl(f^2\d^\mu\varphi\bigr ) \bigr]\nonumber\\ & = \left(t+\frac{\mu n}s\right)ds + \frac ns u^\mu\d_\mu\varphi\,ds - u^\mu\d_\mu\varphi\,dn - \d_\mu\varphi\,d\bigl(f^2\d^\mu\varphi\bigr ) \label{dl2}\end{aligned}\ ] ] where we have used @xmath266 , @xmath267 . now from the josephson equation @xmath268 one sees immediately that ( [ dl2 ] ) is identical to ( [ dl1 ] ) . thus , eq . ( [ ckdlambda ] ) is verified . the last equation that has to be checked is eq . ( [ ckeq4 ] ) . we first write @xmath269 where @xmath270 has been used . we expand @xmath271 \nonumber\\ & = -\d_\mu\bigl[(ts+\mu n)u^\mu u_\nu\bigr ] + nu^\mu\d_\mu\d_\nu\varphi - \d_\mu\bigl(f^2\d^\mu\varphi)\d_\nu\varphi\label{dstheta}\end{aligned}\ ] ] where eq . ( [ eq : jmu - conservation ] ) has been used ; @xmath272\nonumber\\ & = s\d_\nu t + n\d_\nu\mu + s\mu\d_\nu\left(\frac ns\right ) + su^\mu \d_\mu\varphi \d_\nu\left(\frac ns\right ) + nu^\mu \d_\mu\d_\nu\varphi \ . \label{sdtheta}\end{aligned}\ ] ] combining eqs . ( [ dstheta ] ) and ( [ sdtheta ] ) , using the josephson equation , one finds @xmath273 - \d_\mu\bigl(f^2\d^\mu\varphi)\d_\nu\varphi- s \d_\nu t - n\d_\nu \mu\ .\ ] ] it is easy to check that the right hand side is equal to @xmath274 up to an overall sign , with @xmath275 defined in eq . ( [ eq : tmunu - definition ] ) , and hence is equal to zero . j. maldacena , `` the large n limit of superconformal field theories and supergravity , '' adv . theor . math . * 2 * , 231 ( 1998 ) [ int . j. theor . phys . * 38 * , 1113 ( 1999 ) ] [ arxiv : hep - th/9711200 ] . s. s. gubser , i. r. klebanov , and a. m. polyakov , `` gauge theory correlators from non - critical string theory , '' phys . b * 428 * , 105 ( 1998 ) [ arxiv : hep - th/9802109 ] . e. witten , `` anti - de sitter space and holography , '' adv . * 2 * , 253 ( 1998 ) [ arxiv : hep - th/9802150 ] . kovtun , d. t. son , and a. o. starinets , `` viscosity in strongly interacting quantum field theories from black hole physics , '' phys . * 94 * , 111601 ( 2005 ) [ arxiv : hep - th/0405231 ] ; a. buchel and j. t. liu , `` universality of the shear viscosity in supergravity , '' phys . rev . lett . * 93 * , 090602 ( 2004 ) [ arxiv : hep - th/0311175 ] ; g. policastro , d. t. son , and a. o. starinets , `` shear viscosity of strongly coupled @xmath1 supersymmetric yang - mills plasma , '' phys . lett . * 87 * , 081601 ( 2001 ) [ arxiv : hep - th/0104066 ] . p. romatschke and u. romatschke , `` viscosity information from relativistic nuclear collisions : how perfect is the fluid observed at rhic ? , '' lett . * 99 * , 172301 ( 2007 ) [ arxiv:0706.1522 [ nucl - th ] ] . m. luzum and p. romatschke , `` conformal relativistic viscous hydrodynamics : applications to rhic , '' arxiv:0804.4015 [ nucl - th ] . r. baier , p. romatschke , d. t. son , a. o. starinets , and m. a. stephanov , `` relativistic viscous hydrodynamics , conformal invariance , and holography , '' jhep * 0804 * , 100 ( 2008 ) [ arxiv:0712.2451 [ hep - th ] ] . s. bhattacharyya , v. e. hubeny , s. minwalla , and m. rangamani , `` nonlinear fluid dynamics from gravity , '' jhep * 0802 * , 045 ( 2008 ) [ arxiv:0712.2456 [ hep - th ] ] . j. erdmenger , m. haack , m. kaminski , and a. yarom , `` fluid dynamics of r - charged black holes , '' arxiv:0809.2488 [ hep - th ] . n. banerjee , j. bhattacharya , s. bhattacharyya , s. dutta , r. loganayagam , and p. surwka , `` hydrodynamics from charged black branes , '' arxiv:0809.2596 [ hep - th ] . s. a. hartnoll , c. p. herzog , and g. t. horowitz , `` building a holographic superconductor , '' phys . * 101 * , 031601 ( 2008 ) [ arxiv:0803.3295 [ hep - th ] ] . d. t. son , `` hydrodynamics of relativistic systems with broken continuous symmetries , '' int . j. mod . a * 16s1c * , 1284 ( 2001 ) [ arxiv : hep - ph/0011246 ] . r. feynman , _ statistical mechanics _ ( benjamin , new york , ny , 1972 ) . k. skenderis , `` lecture notes on holographic renormalization , '' class . * 19 * , 5849 ( 2002 ) [ arxiv : hep - th/0209067 ] . i. r. klebanov and e. witten , `` ads / cft correspondence and symmetry breaking , '' nucl . phys . b * 556 * , 89 ( 1999 ) [ arxiv : hep - th/9905104 ] . c. t. lane , a. fairbank , and w. m. fairbank , second sound in liquid helium ii , " phys . rev . * 71 * , 600 ( 1947 ) . i. m. khalatnikov and v. v. lebedev , `` relativistic hydrodynamics of a superfluid liquid , '' phys . * 91a * , 70 ( 1982 ) ; v. v. lebedev and i. m. khalatnikov , zh . eksp . * 83 * , 1601 ( 1982 ) [ sov . phys . jetp * 56 * , 923 ( 1982 ) ] .	we study a holographic model of a relativistic quantum system with a global @xmath0 symmetry , at non - zero temperature and density . when the temperature falls below a critical value , we find a second - order superfluid phase transition with mean - field critical exponents . in the symmetry - broken phase , we determine the speed of second sound as a function of temperature . as the velocity of the superfluid component relative to the normal component increases , the superfluid transition goes through a tricritical point and becomes first - order .
simulation is inevitable in studying the evolution of complex cellular systems . large cellular array simulations might require long runs on a serial computer . parallel processing , wherein each cell or a group of cells is hosted by a separate processing element ( pe ) , is a feasible method to speed up the runs . the strategy of a parallel simulation should depend on whether the simulated system is synchronous or asynchronous . a _ synchronous _ system evolves in discrete time @xmath1 . the state of a cell at @xmath2 is determined by the state of the cell and its neighbors at @xmath3 and may explicitly depend on @xmath3 and the result of a random experiment . an obvious and correct way to simulate the system synchrony using a parallel processor is simply to mimic it by the executional synchrony . the simulation is arranged in rounds with one round corresponding to one time step and with no pe processing state changes of its cells for time @xmath2 before all pes have processed state changes of their cells for time @xmath3 . an _ asynchronous _ system evolves in continuous time . state changes at different cells occur asynchronously at unpredictable random times . here two questions should be answered : ( a ) how to specify the asynchrony precisely ? and ( b ) how to carry out the parallel simulations for the specified asynchrony ? unlike the synchronous case , simple mimicry does not work well in the asynchronous case . when geman and geman @xcite , for example , employ executional _ physical _ asynchrony ( introduced by different speeds of different pes ) to mimic the model asynchrony , the simulation becomes irreproducible with its results depending on executional timing . such dependence may be tolerable in tasks other than simulation ( @xcite describes one such task , another example is given in @xcite ) . in the task of simulation , however , it is a serious shortcoming as seen in the following example . suppose a simulationist , after observing the results of a program run , wishes to look closer at a certain phenomenon and inserts an additional ` print ' statement into the code . as a result of the insertion , the executional timing changes and the phenomenon under investigation vanishes . ingerson and buvel @xcite and hofmann @xcite propose various reproducible computational procedures to simulate asynchronies in cellular arrays . however no uniform principle has been proposed , and no special attention to developing parallel algorithms has been paid . it has been observed that the resulting cellular patterns may depend on the computational procedure @xcite . two main results of this paper are : ( i ) a definition of a natural class of asynchronies that can be associated with cellular arrays and ( ii ) efficient parallel algorithms to simulate systems in this class . the following properties specify the _ poisson asynchrony _ , a most common member in the introduced class : + + @xmath4arrivals for a particular cell form a poisson point process . + @xmath4arrivals processes for different cells are independent . + @xmath4the arrival rate is the same , say @xmath5 , for each cell . + @xmath4when there is an arrival , the state of the cell instantaneously changes ; the new state is computed based on the states of the cell and its neighbors just before the change ( in the same manner as in the synchronous model ) . the new state may be equal to the old one . + @xmath4the time of arrival and a random experiment may be involved in the computation . + a familiar example of a cellular system with the poisson asynchrony is the ising model @xcite in the continuous time formulation of glauber @xcite . in this model a cell configuration is defined by the spin variables @xmath6 specified at the cells @xmath7 of a two or three dimensional array . when there is an arrival at a cell @xmath7 , the spin @xmath8 is changed to @xmath9 with probability @xmath10 . with probability @xmath11 , the spin @xmath8 remains unchanged . the probability @xmath10 is determined using the values of @xmath8 and neighbors @xmath12 just before the update time . it is instructive to review the computational procedures for ising simulations . first , the ising simulationists realized that the standard procedure by metropolis , rosenbluth , rosenbluth , teller , and teller @xcite could be applied . in this procedure , the evolution of the configuration is simulated as a sequence of one - spin updates : given a configuration , define the next configuration by choosing a cell @xmath7 uniformly at random and changing or not changing the spin @xmath8 to @xmath9 as required . in the original standard procedure time is discrete . time continuity could have been simply introduced by letting the consecutive arrivals form the poisson process with rate @xmath13 , where @xmath14 is the total number of spins ( cells ) in the system . the problem of long simulation runs became immediately apparent . bortz , kalos , and lebowitz @xcite developed a serial algorithm ( the bkl algorithm ) which avoids processing unsuccessful state change attempts , and reported up to a 10-fold speed - up over the straight - forward implementation of the standard model . ogielski @xcite built special purpose hardware for speeding up the processing . the bkl algorithm is serial . attempts were made to speed up the ising simulation by parallel computations ( friedberg and cameron @xcite , creutz @xcite ) . however , in these computations the original markov chain of the continuous time ising model was modified to satisfy the computational procedure . the modifications do not affect the equilibrium behavior of the chain , and as such are acceptable if one studies only the equilibrium . in the cellular models however , the transient behavior is also of interest , and no model revision should be done . this paper presents efficient methods for parallel simulation of the continuous time asynchronous cellular arrays without changing the model or type of asynchrony in favor of the computational procedure . the methods promise unlimited speed - up when the array and the parallel computer are sufficiently large . for the poisson asynchrony case , it is also shown how the bkl algorithm can be incorporated , further contributing to speed - up . for the ising model , presented algorithms can be viewed as exact parallel counterparts to the standard algorithm by metropolis et al . the latter has been known and believed to be inherently serial since 1953 . yet , the presented algorithms are parallel , efficient , and fairly simple . the `` conceptual level '' codes are rather short ( see figures [ fig : a1c1pe ] , [ fig : s1c1pe ] , [ fig : amcgen ] , [ fig : amcpoi ] , and [ fig : genout ] , ) . an implementation in a real programming language given in the appendix is longer , of course , but still rather simple . this paper is organized as follows : section [ sec : model ] presents a class of asynchronies and a comparison with other published proposals . then section [ sec : algo ] describes the new algorithms on the conceptual level . while the presented algorithms are simple , there is no simple theory which predicts speed - up of these algorithms for cellular arrays and parallel processors of large sizes . section [ sec : perf ] contains a simplified computational procedure which predicts speed - ups faster than it takes to run an actual parallel program . the predictions made by this procedure are compared with actual runs and appear to be rather accurate . the procedure predicts speed - up of more than 8000 for the simulation of @xmath15 poisson asynchronous cellular array in parallel by @xmath16 pes . actual speed - ups obtained thus far were : more than 16 on 25 pes of the balance ( tm ) computer and more than 1900 on @xmath0 pes of the connection machine ( r ) . time @xmath3 is continuous . each cell @xmath7 has a state @xmath17 . at random times , a cell is granted a chance to change the state . the changes , if they occur , are instantaneous events . random attempts to change the state of a cell are independent of similar attempts for other cells . the general model consists of two functions : _ time_of_next_arrival ( ) _ and _ next_state ( ) _ . they are defined as follows : given the old state of the cell and the states of the neighbors just before time @xmath3 , @xmath18 , the next_state @xmath19 is @xmath20 where the possibility @xmath21 is not excluded ; and the time @xmath22 of the next arrival is @xmath23 where always @xmath24 . in and , @xmath25 denotes the result of a random experiment , e.g. , coin tossing , @xmath26 denotes the indexed set of states of all the neighbors of @xmath7 including @xmath7 itself . thus , if @xmath27 , then @xmath28 . subscript @xmath29 expresses the idea of ` just before @xmath3 ' , e.g. , @xmath30 . according to , the value of @xmath8 instantaneously changes at time @xmath3 from @xmath31 to @xmath32 . at time @xmath3 , the value of @xmath8 is already new . the ` just before ' feature resolves a possible ambiguity if two neighbors attempt to change their states at the same simulated time . compare now the class of asynchronies defined by with the ones proposed in the literature : ( a ) model 1 in @xcite reads : `` ... the cells iterate randomly , one at a time . '' let @xmath33 be the probability that cell @xmath7 is chosen . then the following choice of law yields this model @xmath34 where @xmath35 is a random number uniformly distributed on ( 0,1 ) , and @xmath36 is the natural logarithm , @xmath37 . for @xmath38 , the asynchrony was called the _ poisson asynchrony _ in section [ sec : intro ] ; it coincides with the one defined by the standard model @xcite , and by glauber s model @xcite for the ising spin simulations . ( b ) model 2 in @xcite assigns `` each cell a period according to a gaussian distribution ... the cells iterate one at a time each having its own definite period . '' while it is not quite clear from @xcite what is meant by a `` definite period '' ( is it fixed for a cell over a simulation run ? ) , the following choice of law yields this model in a liberal interpretation : @xmath39 where @xmath40 if @xmath41 , and @xmath42 is the cumulative function for the gaussian probability distribution with mean @xmath43 and variance @xmath44 . the probability of @xmath45 is small when @xmath46 and is ignored in @xcite if this interpretation is meant . in a less liberal interpretation , @xmath47 for all @xmath7 , and @xmath48 is itself random and distributed according to the gaussian law . this case is even easier to represent in terms of model than the previous one : @xmath49 . ( 3 ) model trivially extends to a synchronous simulation , where the initial state changes arrive at time 0 and then always @xmath50 is identical to 1 . the first model in @xcite is `` to choose a number of cells at random and change only their values before continuing . '' this is a variant of synchronous simulation ; it is substantially different from both models ( a ) and ( b ) above . in ( a ) and ( b ) , the probability is 1 that no two neighbors attempt to change their states at the same time . in contrast , in this model many neighboring cells are simultaneously changing their values . how the cells are chosen for update is not precisely specified in @xcite . one way to choose the cells is to assign a probability weight @xmath33 for cell @xmath7 , @xmath51 , and to attempt to update cell @xmath7 at each iteration , with probability @xmath33 , independent of any other decision . such a method conforms with the law because the method is local : a cell does not need to know what is happening at distant cells . the second model in @xcite changes states of a fixed number @xmath52 of randomly chosen cells at each iteration . if @xmath53 , this method is not local and does not conform with the law . deterministic computers represent randomness by using pseudo - random number generators . thus , equations and are substituted in the computation by equations @xmath54 and @xmath55 respectively , which do not contain the parameter of randomness @xmath25 . this elimination of @xmath25 symbolizes an obvious but important difference between the simulated system and the simulator : in the simulated system , the observer , being a part of the system , does not know in advance the time of the next arrival . in contrast , the simulationist who is , of course , not a part of the simulated system , can know the time of the next arrival before the next arrival is processed . for example , it is not known in advance when the next event from a poisson stream arrives . however , in the simulation , the time @xmath22 of the next arrival is obtained in a deterministic manner , given the time @xmath3 of the previous arrival : @xmath56 where @xmath5 is the rate , @xmath57 is the @xmath58-th pseudo - random number in the sequence uniformly distributed on @xmath59 , and @xmath60 is the invocation counter . thus , after the previous arrival is processed , the time of the next arrival is already known . if needed , the entire sequence of arrivals can be precomputed and stored in a table for later use in the simulation , so that all future arrival times would be known in advance . + * asynchronous one - cell - per - one - pe algorithm. the algorithm in figure [ fig : a1c1pe ] is the shortest of those presented in this paper . to understand this code , imagine a parallel computer which consists of a number of pes running concurrently . one pe is assigned to simulate one cell . the pe which is assigned to simulate cell @xmath61 , pe@xmath61 , executes the code in figure [ fig : a1c1pe ] with @xmath62 . the pes are interconnected by the network which matches the topology of the cellular array . a pe can receive information from its neighbors . pe@xmath7 maintains state @xmath8 and local simulated time @xmath63 . variables @xmath63 and @xmath8 are visible ( accessible for reading only ) by the neighbors of @xmath7 . time @xmath63 has no connection with the physical time in which the parallel computer runs the program except that @xmath63 may not decrease when the physical time increases . at a given physical instance of simulation , different cells @xmath7 may have different values of @xmath63 . value @xmath64 is a constant which is known to all pes . the algorithm in figure [ fig : a1c1pe ] is very asynchronous : different pes can execute different steps concurrently and can run at different speeds . a statement ` wait_until _ condition _ ' , like the one at step 2 in figure [ fig : a1c1pe ] , does not imply that the _ condition _ must be detected immediately after it occurs . to detect the _ condition _ at step 2 involving local times of neighbors a pe can poll its neighbors one at a time , in any order , with arbitrary delays , and without any respect to what these pes are doing meanwhile . + despite being seemingly almost chaotic , the algorithm in figure [ fig : a1c1pe ] is free from deadlock . moreover , it produces a unique simulated trajectory which is independent of executional timing , provided that : + ( i ) for the same cell , the pseudo - random sequence is always the same , + ( ii ) no two neighboring arrival times are equal . + freedom from deadlock follows from the fact that the cell , whose local time is minimal over the entire array , is always able to make progress . ( this guaranteed worst case performance , is substantially exceeded in an average case . see section [ sec : perf ] . ) the uniqueness of the trajectory can be seen as follows . by ( ii ) , a cell @xmath7 passes the test at step 2 only if its local time @xmath63 is smaller than the local time @xmath65 of any its neighbor @xmath66 . if this is the case , then no neighbor @xmath66 is able to pass the test at step 2 before @xmath7 changes its time at step 4 . this means that processing of the update by @xmath7 is safe : no neighbor changes its state or time before @xmath7 completes the processing . by ( i ) , functions @xmath67 and @xmath68 are independent of the run . therefore , in each program run , no matter what the neighbors of @xmath7 are doing or trying to do , the next arrival time and state for @xmath7 are always the same . it is now clear why assumption ( ii ) is needed . if ( ii ) is violated by two cells @xmath7 and @xmath66 which are neighbors , then the algorithm in figure [ fig : a1c1pe ] does not exclude concurrent updating by @xmath7 and @xmath66 . such concurrent updating introduces an indeterminism and inconsistency . a scenario of the inconsistency can be as follows : at step 3 the _ old _ value of @xmath12 is used to update state @xmath8 , but immediately following step 4 uses the _ new _ value of @xmath12 to update time @xmath63 . in practice , the algorithm in figure [ fig : a1c1pe ] is safe , when @xmath69 for different @xmath7 are independent random samples from a distribution with a continuous density , like an exponential distribution . in this case , ( ii ) holds with probability 1 . unless the pseudo - random number generators are faulty , one may imagine only one reason for violating ( ii ) : finite precision of computer representation of real numbers . + synchronous one - cell - per - one - pe algorithm. if ( ii ) can be violated with a positive probability ( if @xmath3 takes on only integer values , for example ) , then the errors might not be tolerable . in this case the synchronous algorithm in figure [ fig : s1c1pe ] should be used . observe that while the algorithm in figure [ fig : s1c1pe ] is synchronous , it is able to simulate correctly both synchronous and asynchronous systems . two main additions in the algorithm in figure [ fig : s1c1pe ] are : private variables @xmath70 and @xmath71 for temporal storage of updated @xmath72 and @xmath3 , and synchronization barriers ` synchronize ' . when a pe hits a ` synchronize ' statement it must wait until all the other pes hit a ` synchronize ' statement ; then it may resume . two dummy synchronizations at steps 9 and 10 are executed by idling pes in order to match synchronizations at steps 5 and 8 executed by non - idling pes . when ( ii ) is violated , the synchronous algorithm avoids the ambiguity and indeterminism ( which in this case are possible in the asynchronous algorithm ) as follows : in processing concurrent updates of two neighbors @xmath7 and @xmath66 for the same simulated time @xmath73 , first , @xmath7 and @xmath66 read states @xmath74 and times @xmath3 of each other and compute their private @xmath70 s and @xmath71 ( steps 3 and 4 in figure [ fig : s1c1pe ] ) ; then , after the synchronization barrier at step 5 , @xmath7 and @xmath66 write their states and times at steps 6 and 7 , thus making sure that no write interferes with a read . + + aggregation . * in the two algorithms presented above , one pe hosts only one cell . such an arrangement may be wasteful if the communication between pes dominates the computation internal to a pe . a more efficient arrangement is to assign several cells to one pe . for concreteness , consider a two - dimensional @xmath75 array with periodic boundary conditions . let @xmath58 be a multiple of @xmath76 and @xmath77 pes be available . pe@xmath78 carries @xmath79 subarray @xmath78 , where @xmath80 . ( capital @xmath78 will be used without confusion to represent both the subarray index and the set of cells @xmath7 the subarray comprises , e.g. as in @xmath81 ) a fragment of a square cellular array in an example of such an aggregation is represented in figure [ fig : aggr]@xmath82 , wherein @xmath83 . the neighbors of a cell carried by pe1 are cells carried by pe2 , pe3 , pe4 , or pe5 . pe1 has direct connections with these four pes ( figure [ fig : aggr]@xmath84 ) . given cell @xmath7 in the subarray hosted by pe1 , one can determine with which neighboring pes communication is required in order to learn the states of the neighboring cells . let @xmath85 be the set of these pes . examples in figure [ fig : aggr]@xmath82 : @xmath86 is empty , @xmath87\{pe5 } , @xmath88\{pe3 , pe4}. ) mapping of cells to pes , @xmath84 ) the interconnection among the pes which supports the neighborhood topology among the cells , width=556 ] figure [ fig : amcgen ] presents an aggregated variant of the algorithm in figure [ fig : a1c1pe ] . pe@xmath78 , which hosts subarray @xmath78 , maintains the local time register @xmath89 . pe@xmath90 simulates the evolution of its subarray using the algorithm in figure [ fig : amcgen ] with @xmath91 . each cell @xmath92 is represented in the memory of pe@xmath78 by its current state @xmath8 and its next arrival time @xmath63 . note that unlike the one - cell - per - one - pe algorithm , the @xmath63 does not represent the current local time for cell @xmath7 . instead , local times of all cells within subarray @xmath78 are the same , @xmath89 . @xmath89 moves from one @xmath63 to another in the order of increasing value . three successive iterations of this algorithm are shown in figure [ fig : timlin ] , where the subarray @xmath78 consists of four cells : @xmath93 . circles in figure [ fig : timlin ] represent arrival points in the simulated time . a crossed - out circle represents an arrival which has just been processed , i.e. , steps 3 , 4 , and 5 of figure [ fig : amcgen ] have just been executed , so that @xmath89 has just taken on the value of the processed old arrival time @xmath63 , while the @xmath63 has taken on a new larger value . this new value is pointed to by an arrow from @xmath89 in figure [ fig : timlin ] . it is obvious that always @xmath94 if @xmath92 . local times @xmath89 maintained by different pe@xmath78 might be different . a wait at step 3 can not deadlock the execution since the pe@xmath78 whose @xmath89 is the minimum over the entire cellular array is always able to make a progress . slides along a sequence of @xmath63 s in successive iterations of the aggregated algorithm , width=595 ] assuming property ( ii ) as above , the algorithm correctly simulates the history of updates . the following example may serve as an informal proof of this statement . suppose pe1 is currently updating the state of cell @xmath95 ( see figure [ fig : aggr]@xmath82 ) and its local time is @xmath96 . since @xmath97 , this update is possible because the local time of pe5 , @xmath98 , is currently larger than @xmath96 . at present , pe1 receives the state of @xmath99 from pe5 in order to perform the update . this state is in time @xmath98 , i.e. , in the future with respect to local time @xmath96 . however , the update is correct , since the state of @xmath99 was the same at time @xmath96 , as it is at time @xmath98 . indeed , suppose the state of @xmath99 were to be changed at simulated local time @xmath100 , @xmath101 . at the moment when this change would have been processed by pe5 , the local time of pe1 would have been larger than @xmath100 , and @xmath100 would have been the local time of pe5 . after this processing has supposedly taken place , the local time of pe1 should not decrease . yet at the present it is @xmath96 , which is smaller that @xmath100 . this contradiction proves that the state of @xmath99 can not in fact change in the interval ( @xmath102 ) . in the example in figure [ fig : timlin ] , only one @xmath63 supplies @xmath103 . however , the algorithm in figure [ fig : amcgen ] at step 2 commands to select _ a _ cell not _ the _ cell . this covers the unlikely situation of several cells having the same minimum time . if @xmath104 for different @xmath7 are independent random samples from a distribution with a continuous density , this case occurs with the probability zero . on the other hand , if several cells can , with positive probability , update simultaneously , a synchronous version of the aggregated algorithm should be used instead . to eliminate indeterminism and inconsistency , the latter would use synchronization and intermediate storage techniques . these techniques were demonstrated in the algorithm in figure [ fig : s1c1pe ] and their discussion is not repeated here . for an important special case of * poisson asynchrony in the aggregated algorithm * , the algorithm of figure [ fig : amcgen ] is rewritten in figure [ fig : amcpoi ] . this specialization capitalizes on the additive property of poisson streams , specifically , on the fact that sum of @xmath105 independent poisson streams with rate @xmath5 each is a poisson stream with rate @xmath106 . in the algorithm , @xmath107 ; this @xmath105 is equal to @xmath108 in the special case of partitioning into @xmath79 subarrays . unlike the general algorithm of figure [ fig : amcgen ] , in the specialization in figure [ fig : amcpoi ] neither individual streams for different cells are maintained , nor future arrivals @xmath63 for cells are individually computed . instead , a single cumulative stream is simulated and cells are delegated randomly to meet these arrivals . at step 5 in figure [ fig : amcpoi ] , @xmath109 is an @xmath110-th pseudo - random number in the sequence uniformly distributed in ( 0,1 ) . it follows from the notation that each pe has its own sequence . if this sequence is independent of the run ( which is condition ( i ) above ) and if updates for neighboring cells never coincide in time ( which is condition ( ii ) above ) , then this algorithm produces a unique reproducible trajectory . the same statement is also true for the algorithm in figure [ fig : amcgen ] . however , uniqueness provided by the algorithm in figure [ fig : amcpoi ] is weaker than the one provided by the algorithm in figure [ fig : amcgen ] : if the same array is partitioned differently and/or executed with different number of pes , a trajectory produced by the algorithm in figure [ fig : amcpoi ] may change ; however , a trajectory produced by the algorithm in figure [ fig : amcgen ] is invariant for such changes given that each cell @xmath7 uses its own fixed pseudo - random sequence . + * efficiency of aggregated algorithms. both many - cells - per - one - pe algorithms in figure [ fig : amcgen ] and figure [ fig : amcpoi ] are more efficient than the one - cell - per - one - pe counterparts in figure [ fig : a1c1pe ] and figure [ fig : s1c1pe ] . this additional efficiency can be explained in the example of the square array , as follows : in the algorithms in figure [ fig : a1c1pe ] and figure [ fig : s1c1pe ] , a pe may wait for its four neighbors . however , in the algorithms in figure [ fig : amcgen ] and figure [ fig : amcpoi ] , a pe waits for at most two neighbors . for example , when the state of cell @xmath111 in figure [ fig : aggr]@xmath82 is updated , pe1 might wait for pe3 and pe4 . moreover , for at least @xmath112 cells @xmath7 out of @xmath108 , pe1 does not wait at all , because @xmath113 . the cells @xmath7 such that @xmath113 form the dashed square in figure [ fig : aggr]@xmath82 . this additional efficiency becomes especially large if , instead of set @xmath114 in the original formulation of the model , one uses sets @xmath115 or , more generally , @xmath116-th degree neighborhood , @xmath117 . the latter is defined for @xmath118 inductively @xmath119 where @xmath120 for a set @xmath121 of cells is defined as @xmath122 . it is easy to rewrite the algorithms in figure [ fig : a1c1pe ] and figure [ fig : s1c1pe ] for the case @xmath118 . the obtained codes have low efficiency however . for example , in the square array case , one has @xmath123 . thus , if @xmath124 , a cell might have to wait for 12 cells in order to update . in the same example , if one pe carries an @xmath79 subarray , and @xmath125 , then the pe waits for at most three other pes no matter how large the @xmath116 is . moreover , if @xmath126 then in @xmath127 cases out of @xmath108 the pe does not wait at all . + the bkl algorithm * @xcite was originally proposed for ising spin simulations . it was noticed that the probability @xmath10 to flip @xmath8 takes on only a finite ( and small ) number @xmath128 of values @xmath129 , each corresponding to one or several combinations of old values of @xmath8 and neighboring spins @xmath12 . thus the algorithm splits the cells into @xmath128 pairwise disjoint classes @xmath130 , @xmath131, ... @xmath132 . the rates @xmath133 of changes ( not just of the attempts to change ) for all @xmath134 are the same . at each iteration , the bkl algorithm does the following : + _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \(a ) selects @xmath135 at random according to the weights @xmath136 , @xmath137 , and selects a cell @xmath138 uniformly at random . + ( b ) flips the state of the selected cell , @xmath139 . + ( c ) increases the time by @xmath140 , where @xmath141 is a pseudo - random number uniformly distributed in ( 0,1 ) . + ( d ) updates the membership in the classes . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ if the asynchrony law is poisson , the idea of the bkl algorithm can be applied also to a deterministic update . here the probability @xmath10 of change takes on just two values : + @xmath142 if @xmath143 , and @xmath144 if @xmath145 . + accordingly , there are two classes : @xmath146 , the cells which are not going to change and @xmath130 , the cells which are going to change . as with the original bkl algorithm , a substantial overhead is required for maintaining an account of the membership in the classes ( step ( d ) ) . the bkl algorithm is justified only if a large number of cells are not going to change their states . the latter is often the case . for example , in the conways s synchronous _ game of life _ ( gardner @xcite ) large regions of white cells ( @xmath147 ) remain unchanged for many iterations with very few black cells ( @xmath148 ) . one would expect similar behavior for an asynchronous version of the game of life . the basic bkl algorithm is serial . to use it on a parallel computer , an obvious idea is to run a copy of the serial bkl algorithm in each subarray carried by a pe . such a procedure , however , causes roll - backs , as seen in the following example : suppose pe1 is currently updating the state of cell @xmath95 ( figure [ fig : aggr]@xmath82 ) and its local time is @xmath96 , while the local time of pe5 , @xmath98 , is larger than @xmath96 . since @xmath99 is a nearest neighbor to @xmath149 , @xmath99 s membership might change because of @xmath95 s changed state . suppose @xmath99 s membership were to indeed change . although this change would have been in effect since time @xmath96 , pe5 , which is responsible for @xmath99 , would learn about the change only at time @xmath150 . as the past of pe5 is not , therefore , what pe5 has believed it to be , interval [ @xmath151 must have been simulated by pe5 incorrectly , and must be played again . this original roll - back might cause a cascade of secondary roll - backs , third generation roll - backs etc . + * a modified bkl algorithm * applies the original bkl procedure only to a subset of the cells , whereas the procedure of the standard model is applied to the remaining cells . more specifically : an additional separate class @xmath146 is defined . unlike other @xmath152 , @xmath153 , class @xmath146 always contains the same cells . steps ( a ) - ( d ) are performed as above with the following modifications : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \1 ) the weight of @xmath146 at step ( a ) is taken to be @xmath154 . + 2 ) if the selected @xmath7 belongs to @xmath146 , then at step ( b ) the state of @xmath7 may or may not change . the probability @xmath10 of change is determined as in the standard model . + 3 ) the time at step ( c ) should be increased by @xmath155 , where @xmath156 is a pseudo - random number uniformly distributed in @xmath59 . + _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ now consider again the subarray carried by pe1 in figure [ fig : aggr]@xmath82 . the subarray can be subdivided into the @xmath157 `` kernel '' square and the remaining boundary layer . if first degree neighborhood , @xmath158 , is replaced with the @xmath116-th degree neighborhood , @xmath117 , then the kernel is the central @xmath159 square , and the boundary layer has width @xmath116 . in figure [ fig : aggr]@xmath82 , the cells in the dashed square constitute the kernel with @xmath160 . to apply the modified bkl procedure to the subarray carried by pe1 , the boundary layer is declared to be the special fixed class @xmath146 . similar identification is done in the other subarrays . as a result , the fast concurrent bkl procedures on the kernels are shielded from each other by slower procedures on the layers . the roll - back is avoided , since state change of a cell in a subarray does not constitute state or membership change of a cell in another subarray . unless the performance of pe1 is taken into account , the neighbors of pe1 can not even tell whether pe1 uses the standard or the bkl algorithm to update its kernel . as the size of the subarray increases , so does both the relative weight of the kernel and the fraction of the fast bkl processing . + * generating the output. consider the task of generating cellular patterns for specified simulated times . a method for performing this task in a serial simulation or a parallel simulation of a synchronous cellular array is obvious : as the global time reaches a specified value , the computer outputs the states of all cells . in an asynchronous simulation , the task becomes more complicated because there is no global time : different pes may have different local times at each physical instance of simulation . suppose for example , one wants to see the cellular patterns at regular time intervals @xmath161 on a screen of a monitor attached to the computer . without getting too involved in the details of performing i / o operations and the architecture of the parallel computer , it would be enough to assume that a separate process or processes are associated with the output ; these processes scan an output buffer memory space allocated in one or several pes or in the shared memory ; the buffer space consists of @xmath149 frames , numbered 0,1, ... ,@xmath162 , each capable of storing a complete image of the cellular array for one time instance . the output processes draw the image for time @xmath163 on the screen as soon as the frame number @xmath164 ( the reminder of the integer division @xmath165 by @xmath149 ) is full and the previous images have been shown . then the frame is flashed for the next round when it will be filled with the image for time @xmath166 and so on . + the algorithm must fill the appropriate frame with the appropriate data as soon as both data and the frame become available . the modifications that enable the asynchronous algorithm in figure [ fig : amcgen ] to perform this task are presented in figure [ fig : genout ] . in this algorithm , variables @xmath167 and @xmath165 are private ( i.e. , local to pe ) and @xmath168 and @xmath169 are constants whose values are the same for all the pes . note that different pes may fill different frames concurrently . if the slowest pe is presently filling an image for time @xmath163 , then the fastest pe is allowed to fill the image for time no later than @xmath170 . an attempt by the fastest pe to fill the image for time @xmath171 will be blocked at step 4 , until the frame number @xmath172 becomes available . thus , the finiteness of the output buffer introduces a restriction which is not present in the original algorithm in figure [ fig : amcgen ] . according to this restriction , the lag between concurrently processed local times can not exceed a certain constant . the exact value of the constant in each particular instance depends on the relative positions of the update times within the @xmath168-slots . in any case , the constant is not smaller than @xmath173 and not larger than @xmath174 . however , even with a single output buffer segment , @xmath175 , the simulation does not become time - driven . in this case , the concurrently processed local times might be within a distance of up to @xmath168 of each other , whereas @xmath168 might be relatively large . no precision of update time representation is lost , although efficiency might degrade when both @xmath168 and @xmath149 become too small , see section [ sec : perf ] . modeling and analysis of asynchronous algorithms is a difficult theoretical problem . strictly speaking , the following discussion is applicable only to synchronous algorithms . however , one may argue informally that the performance of an asynchronous algorithm is not worse than that of its synchronous counterpart , since expensive synchronizations are eliminated . first , consider the synchronous algorithm in figure [ fig : s1c1pe ] . let @xmath14 be the size of the array and @xmath176 be the number of cells which passed the test at step 2 , figure [ fig : s1c1pe ] . the ratio of useful work performed , to the total work expended at the iteration is @xmath177 . this ratio yields the _ efficiency _ ( or _ utilization _ ) at the given iteration . assuming that in the serial algorithm all the work is useful , and that the algorithm performs the same computation as its parallel counterpart , the speed - up of the parallel computation is the average efficiency times the number of pes involved . here the averaging is done with equal weights over all the iterations . in the general algorithms , @xmath178 is determined using the states of the neighbors of @xmath7 . however , in the important applications , such as an ising model , @xmath178 is independent of states . the following assessment is valid only for this special case of independence . here the configuration is irrelevant and whether the test succeeds or not can be determined knowing only the times at each iteration . this leads to a simplified model in which only local times are taken into account : at an iteration , the local time of a cell is incremented if the time does not exceed the minimum of the local times of its neighbors . a simple ( serial ) algorithm which updates only local times of cells @xmath63 according to the rules formulated above was exercised for different array sizes @xmath58 and three different dimensions : for an @xmath58-element circular array , an @xmath75 toroidal array , and for @xmath179 array with periodic boundary conditions . two types of asynchronies are tried : the poisson asynchrony for which @xmath180 is distributed exponentially , and the asynchrony for which @xmath180 is uniformly distributed in ( 0,1 ) . in both cases , random time increments for different cells are independent . the results of these six experiments are given in figure [ fig : perf1t1 ] . each solid line in figure [ fig : perf1t1 ] is enclosed between two dashed lines . the latter represent 99.99% student s confidence intervals constructed using several simulation runs , that are parametrically the same but fed with different pseudo - random sequences . in figure [ fig : perf1t1 ] , for each array topology there are two solids lines . the poisson asynchrony always corresponds to the lower line . the corresponding limiting values of performances ( when @xmath58 is large ) are also shown near the right end of each curve . for example , the efficiency in the simulation of a large @xmath75 array with the poisson asynchrony is about 0.121 , with the other asynchrony , it is about 0.132 . no analytical theory is available for predicting these values or even proving their separation from zero when @xmath181 . it follows from figure [ fig : perf1t1 ] that replacing exponential distribution of @xmath50 with the uniform distribution results in efficiency increase from 0.247 to 0.271 for a large @xmath58-circle ( @xmath181 ) . the efficiency can be raised even more . if @xmath182 , where @xmath141 is distributed uniformly in ( 0,1 ) , then in the limit @xmath181 , with the student s confidence 99.99% , the efficiency is @xmath183 . it is not known how high the efficiency can be raised this way ( degenerated cases , like a synchronous one , in which the efficiency is 1 , are not counted ) . an efficiency of 0.12 means the speed - up of @xmath184 ; for @xmath185 this comes to more than 1900 . this assessment is confirmed in an actual full scale simulation experiment performed on @xmath186 pes of a connection machine ( r ) ( a quarter of the full computer ) . this simd computer appears well - suited for the synchronous execution of the one - cell - per - one - pe algorithm in figure [ fig : s1c1pe ] on a toroidal array , poisson asynchrony law . since an individual pe is rather slow , it executes several thousand instructions per second , and its absolute speed is not very impressive : it took roughly 1 sec . of real time to update all @xmath187 spins when the traffic generated by other tasks running on the computer was small ( more precise measurement was not available ) . this includes about @xmath188 rounds of the algorithm , several hundred instructions of one pe per round . the 12% efficiency in the one - cell - per - one - pe experiments could be greatly increased by aggregation . the many - cells - per - one - pe algorithm in figure [ fig : amcpoi ] is implemented as a @xmath78 language parallel program for a balance ( tm ) computer , which is a shared memory mimd bus machine . the @xmath75 array was split into @xmath79 subarrays , as shown in figure [ fig : aggr ] , where @xmath58 is a multiple of @xmath76 . because the computer has 30 pes , the experiments could be performed only with @xmath189 , and 25 pes for different @xmath58 and @xmath76 . along with these experiments , a simplified model , similar to the one - cell - per - one - pe case , was run on a serial computer . in this model , quantity @xmath190 is maintained for each pe , @xmath191 . the update of @xmath192 is arranged in rounds , wherein each @xmath192 is updated as follows : + ( i ) with probability @xmath193 , pe@xmath78 updates @xmath192 : @xmath194 where @xmath141 and @xmath36 are the same as in step 5 in figure [ fig : amcpoi ] . here @xmath195 is the probability that the pe chooses a cell @xmath7 so that @xmath196 ; + ( ii ) with probability @xmath197 , the pe must check the @xmath198 of one of its four neighbors @xmath199 before making the update . the @xmath199 is chosen uniformly at random among the four possibilities . if @xmath200 , then @xmath192 gets an increment according to ; otherwise , @xmath192 is not updated . here @xmath201 is the probability that pe will choose a cell @xmath7 in an edge but not in a corner , so that @xmath202 + ( iii ) with the remaining probability @xmath203 , the pe checks @xmath198 and @xmath204 of two of its adjacent neighbors ( for example in figure [ fig : aggr ] , neighbors pe2 and pe3 can be involved in the computation for pe1 ) . the two neighbors are chosen uniformly at random from the four possibilities . again , if both @xmath205 and @xmath206 , then @xmath192 gets an increment according to ; otherwise , @xmath192 is not updated . here @xmath207 is the probability to choose a cell @xmath7 in a corner , so that @xmath208 . as in the previous case , this simplified model simulates a possible but not obligatory synchronous timing arrangement for executing the real asynchronous algorithm . figure [ fig : perfmt1 ] shows excellent agreement between actual and predicted performances for the aggregated ising model . the efficiency presented in figure [ fig : perfmt1 ] is computed as @xmath209 the parallel speed - up can be found as efficiency@xmath210number of pes . for 25 pes simulating a 120@xmath211120 ising model , efficiency is 0.66 ; hence , the speed - up is greater than 16 . for the currently unavailable sizes , when @xmath16 pes simulate a @xmath212 array , the simplified model predicts an efficiency of about 0.8 and a speed - up of about 8000 . in the experiments reported above , the lag between the local times of any two pes was not restricted . as discussed in section [ sec : algo ] , an upper bound on the lag might result from the necessity to produce the output . to see how the bound affects the efficiency , one experiment reported in figure [ fig : perfmt1 ] , is repeated with various finite values of the lag bound . in this experiment , an @xmath75 array is simulated and one pe carries an @xmath79 subarray , where @xmath213 and @xmath214 . the results are presented in figure [ fig : perfbl ] . in figure [ fig : perfbl ] , the unit of measure for a lag is the expectation of time intervals between consecutive arrivals for a cell . for lag bounds greater than 16 , degradation of efficiency is almost unnoticeable , when compared with the base experiment where lag@xmath215 . substantial degradation starts at about 8 ; for the unity lag bound , the efficiency is about half that of the base experiment . however , even for lag bound 0.3 , the simulation remains practical , with an efficiency of about 0.1 ; since 1024 pes execute the task , this efficiency means a speed - up of more than 100 . this paper demonstrates an efficient parallel method for simulating asynchronous cellular arrays . the algorithms are quite simple and easily implementable on appropriate hardware . in particular , each algorithm presented in the paper can be implemented on a general purpose asynchronous parallel computer , such as the currently available bus machines with shared memory . the speed of such implementation depends on the speed of pes and the efficiency of the communication system . a crucial condition for success in such implementation is the availability of a good parallel generator of pseudo - random numbers . to assure reproducibility , each pe should have its own reproducible pseudo - random sequence . the proposed algorithms present a number of challenging mathematical problems , for example , the problem of proving that efficiency tends to a positive limit when the number of pes increases to infinity . + acknowledgments. + i acknowledge the personnel of the thinking machine corp . for their kind invitation , and help in debugging and running the parallel lisp program on one of their computers . particularly , the help of mr . gary rancourt and mr . bernie murray was invaluable . also , i thank andrew t. ogielski and malvin h. kalos for stimulating discussions , debasis mitra for a helpful explanation of a topic in markov chains , and brigid moynahan for carefully reading the text . mmmm s. geman and d. geman , stochastic relaxation , gibbs distributions , and the bayesian restoration of images , _ ieee transactions on pattern analysis and machine intelligence _ , * pami-6 * , 6 , ( novem . 1984 ) , 721741 . # define shared_mem_size ( sizeof(double)10000 ) # define end_time 1000 . # define a 24 / side of small square a pe takes care of/ # define m 5 / number of pes along a side of the big square/ shared int npes = mm , spin[ma][ma ] ; shared float time[m][m ] ; /local times on subarrays/ shared float prob[10 ] ; / * probabilities of state change * / shared float j = 1 . , h = 0 . ; / * energy= -j sum spin spin ' - h sum spin * / shared float t = 1 . ; / * temperature * / shared int ato2 = aa ; shared int am = am ; / * compute flip probabilities * / for ( i = 0 ; i < 5 ; i++ ) for ( j = 0 ; j < 2 ; j++ ) { index = i + 5j ; / index = 0,1, ... ,9 * / my_spin = 2j - 1 ; sum_nei = 2i - 4 ; d_e = 2.(j my_spin * sum_nei + h * my_spin ) ; x = exp(-d_e / t ) ; prob[index ] = x/(1.+x ) ; / * printf("prob[%d]=%f\n",index , prob[index ] ) ; * / } ; / * initialize spins at random , in seedran(seed , b ) , b is dummy/ seedran(31234,1 ) ; for ( i = 0 ; i < ma ; i++ ) for ( j = 0 ; j < ma ; j++ ) { bit = 2frand(1 ) ; / * bit becomes 0 or 1 * / spin[i][j ] = 2bit - 1 ; / spin becomes -1 or 1 * / / * printf("spin[%d][%d]=%d\n",i , j , spin[i][j ] ) ; * / } ; for ( child_id = 0 ; child_id < npes ; child_id++ ) if ( fork ( ) = = 0 ) { tmp_affinity(child_id ) ; / * fixing a pe for process child_id * / work(child_id ) ; / * starting a child pe process * / exit(0 ) ; } work(my_id ) int my_id ; { int i , j ; int coord , var ; int x , y , my_i , my_j , sum_nei , nei_i , nei_j ; int up_i , down_i , left_j , right_j ; int i_base , j_base ; int index ; double frand ( ) ; double r ; double end_time ; while(time[my_i][my_j ] < end_time ) { r = frand(my_id ) ; /pe my_id obtains next pseudo - random number from its own sequence/ x = ra ; y = ( ra - x)a ; /pick a random cell with internal address ( x , y ) within the aa square/ /compute sum of neighboring spins/ sum_nei = 0 ; for ( coord = 0 ; coord < 2 ; coord + = 1 ) for ( var = -1 ; var < 2 ; var + = 2 ) { nei_i = x ; nei_j = y ; if(coord = = 0 ) nei_i + = var ; if(coord = = 1 ) nei_j + = var ; if(0 < = nei_i & & nei_i < a & & 0 < = nei_j & & nei_j < a ) { nei_i + = i_base ; nei_j + = j_base ; } else { / * 4 possible reasons to wait for a neighboring pe * / if(-1 = = nei_i ) while ( time[down_i][my_j ] < time[my_i][my_j ] ) ; if(-1 = = nei_j ) while ( time[my_i][left_j ] < time[my_i][my_j ] ) ; if(nei_i = = a ) while ( time[up_i][my_j ] < time[my_i][my_j ] ) ; if(nei_j = = a ) while ( time[my_i][right_j ] < time[my_i][my_j ] ) ;	a definition for a class of asynchronous cellular arrays is proposed . an example of such asynchrony would be independent poisson arrivals of cell iterations . the ising model in the continuous time formulation of glauber falls into this class . also proposed are efficient parallel algorithms for simulating these asynchronous cellular arrays . in the algorithms , one or several cells are assigned to a processing element ( pe ) , local times for different pes can be different . although the standard serial algorithm by metropolis , rosenbluth , rosenbluth , teller , and teller can simulate such arrays , it is usually believed to be without an efficient parallel counterpart . however , the proposed parallel algorithms contradict this belief proving to be both efficient and able to perform the same task as the standard algorithm . the results of experiments with the new algorithms are encouraging : the speed - up is greater than 16 using 25 pes on a shared memory mimd bus computer , and greater than 1900 using @xmath0 pes on a simd computer . the algorithm by bortz , kalos , and lebowitz can be incorporated in the proposed parallel algorithms , further contributing to speed - up .
the study of ultra - relativistic heavy ion collision has entered a new era with the successful commissioning of the relativistic heavy ion collider at brookhaven . this provides an opportunity to verify the possible occurrence of a phase transition from hadronic matter to deconfined quark matter , where partons are the basic degrees of freedom . at the ( lower ) sps energies , an enhanced production of strangeness , considered to be one of the more robust signatures of quark - hadron phase transition has already been observed @xcite . the initial temperatures likely to be attained at rhic and the lhc are expected to be much larger . a natural question would now be : how quickly is the strangeness equilibrated , if at all ? in heavy - ion collisions strangeness is produced abundantly through the partonic interactions if the temperature @xmath0 mev , the mass threshold of this semi - heavy flavour . the extent of its equilibration would however depend upon such details as the thermal and chemical evolution of the partonic system and the life - time of the hot deconfined phase . it has recently been shown that the chemical equilibration of the light flavours and the gluons slows down due to the radial expansion@xcite . it should then be expected that the extent of strangeness - equilibration can also be affected if allowances are made for the likely radial expansion of the plasma . the present work attempts to get answers to this and related questions . we limit our discussions to the production and evolution of strangeness during the deconfined phase , whose initial conditions are taken from the self screened parton cascade ( sspc ) model @xcite , which has formed the basis of a large number of related studies in recent times . an early work in this direction used the initial conditions obtained from the hijing model and considered only a longitudinal expansion @xcite . in the following we closely follow this treatment and extend it to include transverse expansion as well . our paper has been organized as follows : section ii describes briefly the basic equations of the hydrodynamic and chemical evolution of the partonic gas through the partonic reactions in a ( 1 + 1 ) dimensional longitudinal expansion and a ( 3 + 1 ) dimensional transverse expansion . a brief summary is given in section iii . we start with the assumption that the early ( semi)hard collisions among partons produce a thermalized partonic plasma . the high @xmath1 partons produced early in the collision , then provide a colour screened environment for the production of partons having low @xmath1 and the high density of the partons launches the landau pomeranchuk migdal ( lpm ) suppression mechanism to eliminate the collinear singularity in parton fragmentation . this leads to the the so - called self screened parton cascade model @xcite and can be used to provide plausible initial conditions for the system . the subsequent chemical equilibration is then attained through reactions of the type @xmath2 and @xmath3 . the evolution of the system is now controlled by the equation for conservation of energy and momentum of an ideal fluid : @xmath4 where @xmath5 is the energy density and @xmath6 is the pressure measured in the rest frame of the fluid . the four - velocity vector @xmath7 of the fluid satisfies the constraint @xmath8 . we assume that the distribution functions for partons can be scaled through equilibrium distributions as @xmath9 where @xmath10 is the be ( fd ) distribution for gluons ( quarks ) , and @xmath11 ( @xmath12 ) are the nonequilibrium fugacities , @xmath13 , and @xmath14 is the mass of the parton . now one can write the number density , energy density and pressure for a partially equilibrated multi - component partonic plasma @xcite @xmath15 where @xmath16 and @xmath17 and @xmath18 , \nonumber \\ b_3(x_i ) & = & { { \sum_{n=1}^\infty ( -1)^{n+1 } { 1 \over { ( n x_i)^2 } } k_2(n x_i ) } \over { \sum_{n=1}^\infty ( -1)^{n+1 } \left [ { 1 \over { ( n x_i)^2 } } k_2(n x_i ) + { 1 \over 3 } { 1 \over { ( n x_i ) } } k_1(n x_i ) \right ] } } , \nonumber \\ \label{a3}\end{aligned}\ ] ] for @xmath19 . we also have the colour and spin degeneracy @xmath20 and @xmath21 , where @xmath22 is the mass of the quark and @xmath23 s are modified bessel functions . we take strange quark mass as 150 mev . for massless quarks these expressions simplify considerably and we have @xmath24 , @xmath25 and @xmath26 . we further assume that @xmath27 , which should be valid for negligible net - baryonic density . this should be a reasonable approximation at rhic and lhc energies . the speed of sound ( @xmath28 ) can be obtained from @xmath29 we found it to be close to @xmath30 , as we confine ourselves to @xmath31 200 mev . we must add that several lattice qcd evaluations suggest @xcite that @xmath32 so that @xmath33 . for such a situation the cooling of the plasma would be slower @xcite and thus a much larger time would be available for equilibration . we solve the hydrodynamic equations ( [ hydro ] ) with the assumption that the system undergoes a boost invariant longitudinal expansion along the @xmath34-axis and a cylindrically symmetric transverse expansion @xcite . it is then sufficient to solve the problem for @xmath35 . the chemical equilibration of the species @xmath36 is governed by the master equation@xmath37 where @xmath38 are the rates which propel the system towards chemical equilibration . the system would be in chemical equilibrium when @xmath39 so that @xmath40 . as mentioned earlier , the dominant chemical reactions through which the chemical equilibration proceed @xcite are @xmath3 and @xmath41 . radiative processes involving quarks have substantially smaller cross sections in perturbative qcd , and quarks are less abundant than gluons in the initial phase of the chemical evolution of the parton gas . other elastic scattering processes ensure maintenance of thermal equilibrium . under these assumptions the master equations for different species become @xcite @xmath42 in an obvious notation . the gain and loss term for the gluon fusion process , @xmath43 can be written as @xmath44 where we have used the unitary relation @xmath45 . the above integral can be written as free space cross section for fusion process folded with the distributions for initial particles as @xmath46 { \tilde f}_g(p_1){\tilde f}_g(p_2 ) \ , \label{fusrate1}\end{aligned}\ ] ] with the cross section given by @xmath47 where @xmath48 , the relative velocity between the initial particles . following ref . @xcite the eq.([fusrate1 ] ) can be written @xmath49 so that the rates given in eq.([fusrate ] ) become @xmath50 the net rate for the process @xmath51 can be written as @xmath52 similarly the integral in eq.([gm0 ] ) can also be written in factorised form @xcite @xmath53 the density weighted rates in eqs.([fg],[gm1 ] ) are defined as @xmath54 where the thermally averaged and velocity weighted cross section are @xmath55 where superscript @xmath56 in @xmath57 and @xmath58 denotes to the fusion process for a given flavour . we give the details of these calculations in the next sub - section . in calculating the rates , @xmath59 and @xmath60 , we have also considered the temperature dependence of the strong coupling constant as @xmath61 with the cut - off @xmath62 mev . the scale , @xmath63 , is derived by comparing the inverse of bare gluon propagator in the imaginary time formalism with the static longitudinal propagator @xcite . eq.([master1 ] ) can now be simplified @xmath64 where @xmath65 is the transverse velocity and @xmath66 . if we assume the system to undergo a purely longitudinal boost - invariant expansion , eq . ( [ hydro ] ) reduces to the well - known relation @xcite @xmath67 where @xmath68 is the proper time . using eqs . ( [ eos ] ) in eq . ( [ eq12 ] ) , one can obtain the ultrarelativistic equation of motion @xcite as @xmath69 \varepsilon_g + \sum_i \left [ { { \dot{\lambda}_{i } } \over \lambda_{i } } + 4 { { \dot t } \over t } { \cal d}_\varepsilon(x_i ) \right . \ \ \ \ \ \ \ \ & & \nonumber \\ \left . + { 1 \over \tau } \left ( 1 + { 1\over 3 } b_3(x_i ) \right ) \right ] \varepsilon_i = 0 \ \ & & \label{eq13}\end{aligned}\ ] ] where @xmath70 \right\ } \times \left\ { \sum_{n=1}^\infty ( -1)^{n+1 } \right . \nonumber \\ & \times & \left . \left [ { 1 \over { ( n x_i)^2 } } k_2(n x_i ) + { 1 \over 3 } { 1 \over { ( n x_i ) } } k_1(n x_i ) \right ] \right\}^{-1 } \label{a5}\end{aligned}\ ] ] which reduces to @xmath71 for massless quarks ( antiquarks ) . now the master equations ( [ maschem ] ) can be written as @xmath72 and @xmath73 where @xmath74 } \over { \sum_{n=1}^\infty ( -1)^{n+1 } { 1 \over { ( n x_i ) } } k_2(n x_i ) } } , \label{a6}\ ] ] and @xmath75 , for massless quarks . we can easily see that eqs.([mastf ] ) reduce to eqs.([eq14],[eq15 ] ) if there is no transverse expansion so that the radial velocity , @xmath76 . now one can study evolution of the multi - component partonic plasma in terms of the parton fugacities by solving the eq . ( [ eq15 ] ) for longitudinally and eq . ( [ mastf ] ) for transversely expanding plasma with the parton chemical equilibration rates . in this subsection we briefly recall @xcite the evaluation of equilibration rates for the gluon fusion process ( @xmath77 ) and gluon multiplication process ( @xmath78 ) . the differential cross section given in eq.([fuscsec ] ) can be written as @xmath79 where @xmath80 , which reduces to @xmath81 , for massless incident particles . the matrix element for the process , @xmath82 , can be obtained in terms of mandelstam variable from ref . @xcite as @xmath83 \ \ , \label{matrix}\end{aligned}\ ] ] in which @xmath84 is the current quark mass . however , for massless quarks ( @xmath85 @xmath86 and @xmath87 ) the above eq.([matrix ] ) reduces to @xmath88 \ . \label{matrix0}\end{aligned}\ ] ] this together with eq.([dsdt ] ) diverges logarithmically as @xmath89 . we assume that this logarithmic divergence for massless quarks can be regularized by assigning them mass given by the thermal mass @xcite @xmath90\frac{4\pi}{9}\alpha_s t^2 \ , \label{thmass}\end{aligned}\ ] ] for @xmath91-quark no such approximation is necessary and we use eq.([matrix ] ) with @xmath92 mev . now integrating the matrix element in eq.([dsdt ] ) over the variable @xmath93 , between the limits @xmath94 \ \ , \label{limit}\ ] ] one obtains the total cross - section @xmath95,\end{aligned}\ ] ] where @xmath96 . now the thermally averaged , velocity weighted cross section is defined as @xmath97 using @xmath98 and the thermal average of @xmath91 for a pair of gluons , @xmath99 , we get @xmath100 combining eqs.([eq16 ] ) and ( [ sig2 ] ) we finally obtain the gluon fusion rate . following ref . @xcite we also estimate the @xmath101 for gluon multiplication process from the differential cross - section given by @xcite @xmath102 where the first step function includes the approximate lpm suppression of the induced gluon and the second step function accounts for energy conservation . here @xmath103 denotes the transverse momentum , @xmath104 is the rapidity of the radiated gluon and @xmath105 corresponds the momentum transfer in the elastic collisions . the infrared divergence associated with qcd radiation is regularized by the lpm effect . however , eq . ( [ sig3 ] ) is still having infrared singularities in both scattering cross sections and radiation amplitudes associated with the gluon propagator . one can , approximately , control all these singularities using the debye screening mass @xcite @xmath106 now the regularized gluon density distribution induced by a single scattering is @xmath107 } , \ ] ] and the regularized small angle @xmath108 scattering cross section is @xmath109 the mean free path for the elastic scattering is obtained as @xcite @xmath110 integrating the @xmath111 part analytically we get the gluon multiplication rate as @xmath112 where @xmath113 ^ 2 -4xzx_d } } \right . \nonumber \\ & & \left . + \frac{1}{s\lambda_f^2 } \frac { \cosh^{-1 } \left ( \sqrt x \right ) } { \sqrt{\left [ 1 + x(1+z)y_d\right ] ^2 -4xzy_d } } \ \right \ } , \end{aligned}\ ] ] with @xmath114 and @xmath115 . in the following subsection we discuss the chemical evolution of different parton species . as mentioned earlier we shall be using the initial conditions obtained from the self - screened parton cascade model @xcite . even though they are well - known and used in a large number of studies we reproduce them in table - i for the sake of completeness and an easy reference . we have chosen an initial fugacity for the strange quarks as half of that for the light quarks as in ref . this is consistent with the assumption , which is often made , that the number of flavours is @xmath116 if the mass of @xmath91 quark is taken as zero @xcite . we show our results for the longitudinal expansion for rhic ( upper panel ) and lhc ( lower panel ) energies in figure 1 . as the additional parton production consumes energy , the temperature of the partonic plasma drops considerably faster than the ideal bjorken s scaling ( @xmath117 , @xmath118 and @xmath119 , respectively , are initial temperature and time of the parton gas ) represented by the dashed line . we see that like the light quarks and gluon , the production of strange quarks continues till late in the evolution . we further note that the extent of equilibration for the strange quarks in comparison to that for the light quarks ( @xmath120 ) rises rapidly and once the temperature falls below about @xmath121 mev ( @xmath122 ) it gets more or less frozen by this time . thus we conclude that the equilibration of strangeness production may imply the existence of quark matter at a temperature of more than about 300 mev for a time of the order of a few fm/@xmath123 . the other aspects of variation of @xmath124 and @xmath125 have already been discussed by several authors @xcite . we do note that the plasma is not fully equilibrated chemically at either rhic or lhc energies by the time the temperature drops below 200 mev . we do expect that additional quark production may occur due to gluon fragmentations during hadronisation @xcite leading to a chemically equilibrated hot hadronic matter . before giving our results for transverse expansion we need to specify the profile of the initial energy density and the fugacity of the partonic system . following ref . @xcite we take the initial energy density as @xmath126^{1/2 } \theta(r - r),\ ] ] where @xmath127 is the transverse dimension of the system , @xmath128 is the radial distance , and @xmath129 is the `` average '' initial energy density ( see tab . the profile plays an important role in defining the boundary of the hot and dense deconfined matter and affects the transverse expansion through the introduction of pressure gradients . we have further taken @xmath130 as before @xcite . any other variation will require an additional parameter . we give our results for the radial variation of @xmath131 , @xmath132 and @xmath120 in figs . 2 and 3 for rhic and lhc energies , respectively , for various times along the constant energy density contours with @xmath133 . here @xmath134 is defined @xcite through @xmath135 . we see that the fugacities attain their highest values near @xmath136 and rise rapidly first and only slowly later in time . we also see a result unique to chemical evolution with transverse expansion that the fugacities may even start reducing towards the end of the qgp phase when the radial velocity ( gradient ) becomes very large ( see ref . @xcite for explanation ) . we can use our results for the radial variation of the energy density and the fugacities obtained here to estimate the extent of partonic equilibration as a function of the initial energy density . comparing results for this from figures 4 and 5 we note that once the energy density is beyond about 2040 gev , the final fugacities for all the partons increase almost linearly with the square of the energy density . this is a very interesting result . let us identify the `` local '' energy density at the radius @xmath128 with the average energy density attained in a non - central collision or with a central collision involving lighter nuclei . recall that the partonic models suggest that the initial temperature attained in such collisions through the production of minijets varies as @xmath137 @xcite . this should then imply that the extent of equilibration of strangeness produced would rise as @xmath138 , if we can identify @xmath139 with @xmath140 , where @xmath141 is the number of participants . in actual cases the variation may be somewhat modified from this naive expectation due to the considerations of shadowing and jet quenching @xcite . if on the other hand the energy densities etc . are decided by the considerations of parton saturation @xcite then the initial temperature would vary as @xmath142 and the energy density as @xmath143 . the results of figs . 4 and 5 would then imply that the extent of strangeness equilibration should increase as the number of participants , @xmath144 ( see also ref . @xcite for a slight reformulation of the parton saturation model ) . it is also interesting to note that a recent analysis @xcite of centrality dependence of the extent of strangeness equilibration at cern sps energies in pb+pb collisions gives a linear increase for this with the number of participants . rafelski and coworkers have studied the strangeness equilibration for more than two decades in various details . we shall refer to only a recent work by these authors @xcite . in ref . @xcite the following assumptions have been made which are at variance with our works : 1 . the light quarks and gluons are assumed to have attained chemical equilibrium ( @xmath145 ) when the evolution of strangeness begins . a boltzmann approximation is used to describe the phase space distribution of @xmath91-quarks . the flavour changing process , @xmath146 , has been included . we have neglected them assuming that the gluon fusion process @xmath147 dominates in a chemically undersaturated plasma . now in the boltzmann limit the evolution of the @xmath91-quark fugacity can be written from ( [ eq15 ] ) as @xmath148 + \frac{1}{\tau } = r^s_2\frac{n_g}{n_s } \left ( 1-\frac{\lambda_s^2}{\lambda_g^2}\right ) . \end{aligned}\ ] ] this can be rewritten as @xmath149 + \frac{n_s}{\tau } \nonumber \\ & = & r^s_2n_g \left ( 1-\frac{\lambda_s^2}{\lambda_g^2}\right ) , \end{aligned}\ ] ] which corresponds to eq.(11 ) of ref . we see that the @xmath150 and @xmath151 terms in the left hand side here are absent in eq.(11 ) of ref . @xcite , due to the above discussion . one can thus see that the extent of strangeness equilibration attained in the work of ref . @xcite is much larger than in the present work . as a further check , we have repeated our calculations , assuming @xmath152 with the same energy density as before , ( see table i ) and further taking @xmath153 as in the work of ref . we found that now our parameter @xmath154 rises to about 0.4 at rhic energies and up to 0.72 at lhc energies , for the case of longitudinal expansion ( see fig . when the transverse expansion is allowed these numbers reduce to 0.66 at lhc energies and 0.38 at rhic energies , respectively . we thus realize that the extent of equilibration of strangeness depends sensitively on the initial conditions and also on the evolution mechanism . wong has studied @xcite the chemical equilibration of plasma essentially in a formalism similar to that used in the present work , but only with longitudinal expansions . his values for @xmath155 are also much larger . we have studied the evolution and production of strangeness through the partonic interactions in a chemically equilibrating and expanding multi - component partonic gas with the initial conditions obtained from sspc model . we find that most of the strange quarks are produced when the temperature is still more than about 300 mev ( @xmath156 ) . thus a chemically equilibrated plasma is expected to imply the existence of qgp phase for a duration of several fm/@xmath123 . we also find approximately that the extent of strangeness equilibration rises linearly with the square of the initial energy density within our approach . this may help us to obtain the scaling of the initial energy density with the number of participants and distinguish between the minijet and the partonic saturation models of parton production . it is rather interesting that the charged particle multiplicity in au + au collision at @xmath157 gev measured by the phenix collaboration @xcite shows a behaviour which is a superposition of two terms , a linear increase with the number of participants and a linear increase with the number of collisions ( which varies as @xmath158 ) . one of us ( d.p . ) acknowledges the hospitality of the theory group of saha institute of nuclear physics during her short term visit when most of the work was done . we also thank x .- n . wang for useful correspondence . j. rafelski , phys . rep . * 88 * , 331 ( 1982 ) ; p. koch , b. mller and j. rafelski , phys . rep . * 142 * , 167 ( 1986 ) ; m. gyulassy , nucl . a * 418 * , 594@xmath159 ( 1984 ) . t. matsui , b. svetitsky and l. d. mclerran , phys . rev . d34 , 783 ( 1986 ) ; _ ibid _ * 34 * , 2047 ( 1986 ) . k. geiger , phys . d * 48 * , 4129 ( 1993 ) . j. letessier and j. rafelski , nucl . a * 661 * , 497@xmath159 ( 2000 ) ; phys . c * 59 * , 947 ( 1999 ) ; j. phys . g * 25 * , 295 ( 1999 ) . d. k. srivastava , m. g. mustafa , and b. mller , phys . b * 396 * , 45 ( 1997 ) ; d. k. srivastava , m. g. mustafa , and b. mller , phys . c * 56 * , 1064 ( 1997 ) . d. m. elliott and d. h. rischke , nucl . a * 671 * , 583 ( 2000 ) . k. j. eskola , b. mller , and x. n. wang , phys . b * 374 * , 20 ( 1996 ) p. levai and x .- wang , @xmath160hep - ph/9504214@xmath161 . e. laermann , nucl . a * 610 * , @xmath162 ( 1996 ) . d. pal , b. k. patra , and d. k. srivastava , eur . phys . j. c * 17 * , 179 ( 2000 ) . h. von gersdorff , l. mclerran , m. kataja , and p. v. ruuskanen , phys . d * 34 * , 794 ( 1986 ) . t. s. biro , e. van doorn , b. mller , m. h. thoma , and x. n. wang , phys . c * 48 * , 1275 ( 1993 ) . p. levai , b. mller and x .- wang , phys . c * 51 * 3326 ( 1995 ) . see , e.g. , m. h. thoma , @xmath163 application of high temperature field theory " , quark - gluon plasma -2 , pg . 51 , ed . : r. c. hwa , world scientific , 1995 . j. d. bjorken , phys . d * 27 * , 140 ( 1983 ) . b. l. combridge , nucl . b * 151 * , 429 ( 1979 ) . t. s. biro , p. levai and b. mller , phys . d * 42 * , 3078 ( 1990 ) . t. alther and d. seibert , phys . d * 49 * , 1684 ( 1994 ) . r. stock , phys . b456 , 277 ( 1999 ) . j. ellis and k. geiger , phys . rev . d54 , 1967 ( 1996 ) . b. mller , m. g. mustafa and d. k. srivastava , heavy ion physics * 5 * , 394 ( 1997 ) . b. k. patra and d. k. srivastava , phys . b * 505 * , 113 ( 2001 ) . r. c. hwa and k. kajantie , phys . * 56 * , 696 ( 1986 ) . wang , phys . rep . * 280 * , 287 ( 1997 ) . wang and m. gyulassy , phys . * 86 * , 3496 ( 2001 ) . k. eskola , k. kajantie , p. v. ruuskanen and k. touminen , nucl . b * 570 * , 379 ( 2000 ) . d. kharzeev and m. nardi , phys . b * 507 * , 121 ( 2001 ) . j. cleymans , b. kaempfer , and s. wheaton , nucl - th/0110035 . j. rafelski and j. letessier , phys . b469 , 12 ( 1999 ) . s. m. h. wong , phys . c * 58 * , 2358 ( 1998 ) ; _ ibid _ * 56 * , 1075 ( 1997 ) . phenix collaboration , k. adcox _ et al . _ , rev . lett . * 86 * , 3500 ( 2001 ) . .initial conditions for the hydrodynamical expansion phase in central collision of two gold nuclei at bnl rhic and cern lhc energies , respectively , from sspc@xcite given in first and second row . initial conditionis given in third and fourth row are obtained with the same energy densities as before along with @xmath164 . [ cols="<,^,^,^,^,^,^ " , ]	we evaluate the strangeness production from equilibrating and transversely expanding quark gluon plasma which may be created in the wake of relativistic heavy ion collisions . we consider boost invariant longitudinal and cylindrically symmetric transverse expansion of a gluon dominated partonic plasma , which is in local thermal equilibrium . initial conditions obtained from the self screened parton cascade model are used . we empirically find that the final extent of the partonic equilibration rises almost linearly with the square of the initial energy density . this along with the corresponding variation with the number of participants may help us distinguish between various models of parton production . epsf -0.5 in 0.4 in 0.4 in 0.4 in 0.25 in pacs : 12.38.mh , 24.85.+p , 25.75.-q , 13.85.qk 0.25 in
there is much interest at present in the possible use of the new generation of high - power laser facilities ( in particular the national ignition facility at livermore and the laser megajoule in bordeaux ) to simulate astrophysical phenomena such as supernovae . at first sight this programme appears to suffer from one obvious drawback . the phenomena one wishes to simulate generally involve _ explosions _ while the laser facilities are designed to produce _ remarkably , as we will show , this is not a problem . under certain , not too restrictive , conditions there exists an exact mathematical duality which allows one to transform an explosion problem to an implosion problem and vice versa . thus it is possible , in a precise sense , to use _ implosion _ experiments to simulate _ exploding _ systems . the euler equations of perfect gas dynamics can be conveniently written in the form @xmath0 where @xmath1 denotes the lagrangian , material or convective derivative defined by = + * u * and @xmath2 is the mass density , @xmath3 the thermal energy density , @xmath4 the pressure and @xmath5 the velocity . in addition to these differential equations one algebraic relation is needed , an equation of state relating the pressure to the mass and energy densities the simplest being a polytropic equation of state , p = ( -1 ) with @xmath6 a constant . these equations are mathematically equivalent to the mass , momentum and energy conservation equations in smooth regions of the flow . only at shocks is it necessary to revert to the fundamental conservation forms to recover the correct shock jump conditions . now consider the following transformation of the dependent and independent variables , @xmath7 where for the moment @xmath8 is an arbitrary function of time . apart from the , at first sight rather strange , time - dependent scaling factors this is essentially a transformation to a coordinate system which is expanding or contracting with a scale factor @xmath8 . if we define @xmath9 and note that = a^2 ddt(1a ) = - d adt it is easy to see that it is an involutionary transformation with inverse obtained by simply interchanging the starred and unstarred quantities @xmath10 let us now consider how the dynamical equations transform under this change of variables . it is easy to see that @xmath11 and thus , after some elementary algebra , @xmath12 remarkably , we see that if the scale factor @xmath13 is such that ^ * = d^2 a^d t^2 = - a^2 a = 0 and the gas is a polytrope of exponent @xmath14 with @xmath15 then the euler equations are _ invariant _ under this transformation . note that , because the euler equations in conservation form are algebraically equivalent to the simplified forms , the conservation forms are also invariant and thus the whole structure of ideal gas dynamics , including the rankine - hugoniot shock relations , is preserved . the condition that the acceleration of the scale factor be zero , @xmath16 , requires that @xmath8 be a linear function of @xmath17 and , without loss of generality , we can take @xmath18 where @xmath19 is a constant characteristic expansion time . the time transformation is then t^ * = = t_0 ^ 2 = - t_0 ^ 2 t and it is convenient to set the constant to zero and choose t^ * = -t_0 ^ 2 t , t = -t_0 ^ 2t^. the initial singularity of the expansion in physical space occurs at @xmath20 and is mapped to @xmath21 , the long term behaviour as @xmath22 is mapped to @xmath23 . it is important to note that in the dual representation the time variable is bounded from _ above _ , @xmath24 , whereas in physical space it is bounded from _ below _ , @xmath25 . the remarkable result is that for an ideal gas of point particles with no internal structure ( which is what the 5/3 polytrope is ) hydrodynamics in a uniformly expanding system is exactly equivalent to hydrodynamics in a static system . this result , or special forms of it , appears to have been discovered a number of times by cosmologists ( where the idea of factoring out the general expansion of the universe is very natural ) ; a recent discussion is that of martel and shapiro ( 1998 ) where they propose the felicitous name of `` supercomoving variables '' to describe this transformation . what does not seem to have been generally noted is that this transformation can be used outside the cosmological context ( however poyet and spiegel , 1979 , did use a variant in an analysis of stellar pulsations ) . the fact that the transformation is exact for the gas of ideal point particles strongly hints that it is derived from a similar result for the free particle motion . in fact there is such a duality , although it is almost trivial . the freely moving point particle moves along a straight line trajectory , = x_0 + v_0 t , with starting point @xmath26 and velocity @xmath27 . if we write this as = v_0 + x_0 1 t we see that there is a dual representation of the trajectory in which @xmath17 is replaced by @xmath28 , lengths are scaled by a factor proportional to time , initial points and final velocities are interchanged , but the trajectory remains a straight line . if collisions are instantaneous , localised and elastic they look the same in either system , and thus in both systems one can write down a boltzmann equation and then derive the hydrodynamic equations as limits of moments of the boltzmann equation . this approach also shows that higher order effects , such as viscosity and heat conduction , can formally be treated in the same way ; however the resulting transformed transport coefficients will in general have unphysical time dependencies ( for an application see drury and stewart , 1976 ) . this analysis also shows that similar results will hold in different numbers of spatial dimensions , but the equation of state will have to correspond to the ideal gas in that number of dimensions . in @xmath29 spatial dimensions it is easy to verify that the `` super - comoving '' transformation takes the form @xmath30 and that the euler equations are invariant if the gas has a polytropic equation of state such that d p = 2 corresponding to an adiabatic exponent = 1 + 2d . an interesting way of looking at this transformation ( for which we are indebted to our colleague etienne parizot ) is that it provides an analogue in spherical geometry to the freedom that galilei transformations allow in planar geometry . if we are looking at a planar shock , it is often convenient to transform to a reference frame where the upstream medium , or the downstream medium , or the shock itself , appears stationary . in spherical systems one can not apply galilei boosts because the origin is fixed , however this transformation , by allowing one to take out an arbitrary uniform expansion , gives one much the same freedom . computational studies of the evolution of a supernova remnant commonly start with initial conditions of dense pressure - free ejecta expanding ballistically away from the site of the explosion , which it is convenient to locate at the coordinate origin , and interacting with a stationary , or slowly moving , ambient medium of much lower density and negligible pressure . to illustrate the application of the duality transformation let us consider the simple , if somewhat artificial , case of uniform density ejecta interacting with a uniform and stationary ambient medium in perfect spherical symmetry . then the initial conditions correspond to @xmath31 in the region @xmath32 occupied by the ejecta ( @xmath33 is the maximum expansion speed of the ejecta ) and @xmath34 in the external ( @xmath35 ) medium of constant density @xmath36 . the sweep - up time @xmath37 corresponds to the point where the ejecta , if expanding unimpeded , would have a density equal to the ambient medium . this defines the physical problem of expanding ejecta interacting with a stationary environment . let us now consider the dual problem obtained by applying the transformation with scale factor a ( t ) = t. then the dependent variables transform as @xmath38 so that the explosion , which occurs at @xmath20 in physical problem , occurs at @xmath39 in the dual problem . conversely the asymptotic evolution as @xmath40 in the physical problem is mapped to the behaviour at @xmath23 in the dual problem . the ejecta density in the dual problem is constant , ^(r^ * , t^ * ) = a^3 ( r , t ) = _ 0 and the velocity is zero , @xmath41 , in @xmath42 . however the ambient medium is now time - dependent with density , in the region @xmath43 ^(r^ , t^ * ) = ( t)^3 _ 0 = ( -t^)^-3 _ 0 and velocity u^(r^ * , t^ * ) = r^t^. thus in the dual problem we have _ stationary _ ejecta interacting with an _ imploding _ ambient medium whereas in the physical problem we have _ exploding _ ejecta interacting with a _ stationary _ ambient medium . instead of the initial explosion at @xmath20 in the physical problem we have the final crunch at @xmath44 in the dual problem . the evolution in physical space of the supernova remnant structure has been often discussed and is well - known ( eg truelove and mckee , 1999 ; dwarkadas and chevalier , 1998 ) . at early times , @xmath45 , the bulk of the ejecta expand ballistically except for a thin interaction region on the outside consisting of a forward shock running into the ambient medium , a zone of hot shocked ambient medium , a contact discontinuity , a zone of shocked ejecta and a reverse shock propagating slowly into the ejecta . at later times , when the mass of swept up ambient material becomes comparable to the ejecta mass , the reverse shock detaches itself from the contact discontinuity and implodes on the centre and the outer forward shock approximates the self - similar sedov solution for a strong point explosion in a cold gas . in the dual system the interaction looks a little different , and in some ways is simpler . initially we have the stationary sphere of high density material ( which for convenience we continue to call the ejecta , although in the dual representation it has not been ejected but is simply sitting there ) surrounded by a very low density converging flow . the inflowing gas has to decelerate at a shock which stands about 10% further out in radius than the edge of the ejecta . writing for convenience @xmath46 there is an exact similarity solution in which @xmath47 , @xmath48 and @xmath49 in the region external to the sphere of ejecta . this steeply rising pressure ( @xmath50 ) drives the reverse shock into the ejecta and starts the implosion of the ejecta . = 0.9 at later times , as the ejecta collapse , the shock in the imploding ambient medium also moves inwards thereby reducing the rate of increase of the pressure . transforming the sedov solution to the dual system we see that the shock radius scales as r^ * ^3/5 and the postshock pressure as p^*^-19/5 . figure 1 attempts to show schematically the relation between the two representations . we note in passing that the dual representation is also useful for analytic and numerical studies ; this aspect will be explored in a companion paper ( dwarkadas and drury , in preparation ) . the perfectly symmetric explosion is neither realistic nor especially interesting ; it is the easiest case to analyse numerically and there is no reason to suppose that a laboratory simulation would yield any additional information . however reality is more complicated . it is clear that the ejecta emerging from real supernova explosions are highly nonuniform on a wide range of scales and that to calculate the resulting remnant evolution in three dimensions is likely to remain a computationally challenging problem for some considerable time ( cf arnett , 1999 ) . the interesting implication of this work is that it should be possible with the new generation of implosion facilities to simulate precisely this problem , the interaction of highly structured ejecta with their surroundings including all the effects of spherical geometry . one can easily imagine constructing a solid target whose density distribution models the density distribution of the expanding ejecta . if this target is then used in an implosion experiment , and if the momentum loading on the surface is tailored to rise in the same manner as the pressure behind the forward shock in the dual system , a steep initial rise as @xmath51 decreasing to @xmath52 , the evolution of the internal structures including all the turbulent mixing , instabilities and shock formation , should be exactly replicated . we emphasise finally that the transformation discussed in this paper is additional to and complements the well - known linear scaling relations as excellently discussed by ryutov et al ( 1999 ) in the astrophysical context , or connor and taylor ( 1977 ) in the plasma physics context . dimensional similarity and scaling arguments are obviously central to any attempt at simulation on a laboratory scale of astrophysical systems , however precisely because they are very general and linear they can not turn an explosion into an implosion . the remarkable nonlinear symmetry discussed in this paper is specific to the ideal gas equation of state , but subject to this constraint gives a powerful new degree of freedom in simulation studies by allowing an arbitrary uniform expansion or contraction to be factored out thereby transforming an explosion problem to an implosion one or vice versa . this work was in part supported by the eu under the tmr programme , contract fmrx - ct98 - 0168 . some of it was carried out while ld was a visitor at the research centre for theoretical astrophysics of the university of sydney . = 0 pt = 5 pt plus 5 pt	the euler equations of ideal gas dynamics posess a remarkable nonlinear involutional symmetry which allows one to factor out an arbitrary uniform expansion or contraction of the system . the nature of this symmetry ( called by cosmologists the transformation to supercomoving variables ) is discussed and its origin clarified . it is pointed out that this symmetry allows one to map an explosion problem to a dual implosion problem and vice versa . the application to laboratory simulations of supernova remnants is considered ; in principle this duality allows the complete three - dimensional evolution of highly structured explosion ejecta to be modelled using a static target in an implosion facility . # 1#1 t # 1#1x # 1d#1d t
extended emission - line regions ( eelr ) around radio galaxies are common , and often aligned with the axes of the extended radio emission ( e.g. mccarthy et al . 1987 ; mccarthy 1993 ; mccarthy , baum & spinrad 1996 ) . there is considerable debate as to whether eelr are ionized by photons from the nucleus or by shocks driven by the jets ( e.g.tadhunter et al . most eelr around low - redshift sources can adequately be described by central - illumination models , while sources which show evidence for shock ionization ( such as extreme emission - line kinematics ) are largely at high redshift . however , there is evidence that a small minority of low - redshift radio galaxies are also more adequately described by a shock - ionization model . if the ionization is due to shocks , the physical conditions in and kinematics of the eelr may give us important information about the energetics of the radio source as a whole . one good candidate low - redshift shock - ionization source is 3c171 , an unusual @xmath0 radio galaxy . its large - scale radio structure has been well studied ( heckman , van breugel & miley 1984 ; blundell 1996 ) . in its inner regions it is similar to a normal frii , but low - surface - brightness plumes instead of normal lobes extend north and south from the hotspots . high - resolution images show knotty jets connecting the radio core with the hotspots ( hardcastle et al . 3c171 is associated with a prominent eelr , elongated along the jet axis . heckman et al . found that the emission - line gas was brightest near the hotspots , suggesting that the radio - emitting plasma is responsible for powering the emission - line regions . more recently clark et al . ( 1998 ) have made wht observations of the emission - line regions and shown that their ionization states are most consistent with a shock - ionization model . the unusually strong eelr in this source may be indicative of an environment unusually rich in dense , warm gas ; the interactions between the jets and this gas might then be responsible for the peculiar large - scale radio structure . heckman et al . pointed out the depolarization of the radio source near the hotspots at frequencies up to 15 ghz . the strong spatial relationship between the eelr ( or the material traced by the eelr ) and depolarization was pointed out by hardcastle et al . ( 1997 ) , who superposed an 8-ghz very large array ( vla ) image on the _ hubble space telescope _ ( _ hst _ ) snapshot ( de koff et al . 1996 ) in the f702w filter ( which at this redshift passes several strong lines ) . the source has since been observed with a number of other _ hst _ filters , and fig . 1 shows a superposition of the 8-ghz map on an _ hst _ image in the [ oiii]495.9/500.7 nm lines . regions of low polarization in the radio map can be seen to correspond to the positions of emission - line material . the eastern hot spot is almost completely unpolarized at 8 ghz and below , while the western hot spot is partially depolarized . the close assocation between the emission - line regions and the radio depolarization in 3c171 opens up the possibility of using radio data to constrain unknown physical conditions , such as the electron number density and the magnetic field strength in the gas surrounding the source . in this paper , i present new radio observations of 3c171 , and discuss the constraints that they place on the nature of its environment . throughout the paper , i use a cosmology with @xmath1 km s@xmath2 mpc@xmath2 , @xmath3 and @xmath4 . at the redshift of 3c171 , 1 arcsec corresponds to 4.07 kpc . the degree of depolarization is in general expected to be reduced at higher frequencies and higher resolutions . as parts of 3c171 s hotspots had been found to be completely depolarized at the full resolution of the vla ( @xmath5 arcsec ) at 8.1 ghz , i elected to observe the source with the a and b arrays of the vla at the u and k bands , @xmath6 and @xmath7 ghz . in addition , i made use of the a - array 8.1-ghz observations discussed by hardcastle et al . ( 1997 ) . details of the observations used are listed in table [ obstab ] . the new observations were carried out in fast - switching mode with an integration time of 3.3 s , using frequent observations of a nearby bright point - source calibrator , 0749 + 540 . the cycle time used was 320 s , with 80 s on calibrator and 240s on source , giving typical times of 50 s on the calibrator and 210 s on source after dead time due to slewing . ( the low efficiency was due to the lack of a good nearby calibrator . ) primary referenced pointing was carried out , using short x - band observations of 0749 + 540 taken every hour , to avoid loss of sensitivity due to vla pointing errors . the observations were made with two nearby observing frequencies ( table [ obstab ] ) with a 50-mhz bandwidth in each . 3c286 was used as the flux and polarization position angle calibrator . the data were reduced in a standard manner within aips . elint was used to correct for the dependence of antenna gain on elevation . the use of fast - switching mode led to observations which were very well phase- and amplitude - calibrated after the standard calibration procedure : however , there was enough signal for a single iteration of phase self - calibration for each dataset , which gave a marginal improvement in image quality . finally , the a- and b - array datasets were combined using dbcon , uvfix was used to align the datasets at the three wavelengths at the radio core , using the k - band dataset as the reference , and final images were made using imagr , combining the data from the two observing frequencies in each band . this resulted in images at effective frequencies of 8.09 , 14.94 and 22.46 ghz . for simplicity , these frequencies will be referred to as 8 , 15 and 22 ghz from now on . .vla observations [ cols= " < , < , > , > , < , > " , ] the ` image ' column describes the resolution of the image used : ` full ' denotes the full resolution of the dataset , ` 8 ' means matched to the 8-ghz dataset , and ` 15 ' means that the image was matched to the 15-ghz dataset ( the images are those described in table [ maps ] ) . all images were made with matched shortest baselines . the errors on the polarized intensity are determined from the off - source r.m.s . noise in the stokes @xmath8 and @xmath9 images . errors on the total intensity are similar in magnitude , but are not quoted as they are not a significant contributor to the errors in fractional polarization . we can begin by making the assumption that the depolarizing medium is external to the source . this is justified by the extremely good , though not perfect , association between the eelr and the depolarization . no model places the eelr inside the radio source it is spatially quite distinct from , and physically larger than , the radio emission at various points ( fig . 1 ) and at the same time it is clearly related to the depolarization . we can therefore treat the depolarizing material as an external screen of dimensions comparable to those of the eelr . the situation is then a classical problem in the theory of radio polarization , discussed by burn ( 1966 ) . if a screen is external , two extreme cases can be identified . in the first , the faraday screen is fully resolved by the telescope beam : in this case , no depolarization should be observed , but faraday rotation should be seen . some sources in rich cluster environments appear to be in this regime , in that the faraday screen is partially or fully resolved ( e.g. perley & carilli 1996 ) . in 3c171 , however , we see significant depolarization without much associated faraday rotation , as fig . [ depol ] shows . this is likely to be due to a largely unresolved foreground screen . in the limit that the size scale @xmath10 of the individual depolarizing regions ( regions of uniform magnetic field ) is much smaller than the resolution element , _ and is well defined _ , and that there are many regions along each line of sight , the degree of polarization @xmath11 of a part of the source seen through such a screen at a given frequency @xmath12 is given by @xmath13 \label{depeq}\ ] ] ( burn 1966 ) where @xmath14 is the intrinsic polarization , @xmath15 is a constant with value @xmath16 rad nt@xmath2 m kpc@xmath2 , @xmath17 is the dispersion in the product of number density and magnetic field strength and @xmath18 is the line - of - sight depth through the medium . we can try to apply this model to the observations , but there are a number of reasons why it should be treated with caution . the first is a practical one : the measurements of the polarized intensity of the hotspots clearly include regions with different degrees of polarization ( fig . [ depol ] ) and so , presumably , different physical conditions in the depolarizing screen . at best we will obtain some emission - weighted average of the physical conditions . to reduce this effect , i consider below only the e hotspot , which exhibits more uniform depolarization . the model is also not realistic . it assumes a single physical size scale for the variations in faraday depth , parametrized by @xmath10 . as laing ( 1984 ) points out , realistic screens will have a range of scale sizes , which tends to wash out the sharp cutoff in polarization at low frequencies . this is probably the case for our observations : the degree of polarization at 8 ghz , though low , is still too high to be fitted with a simple model of the form of eq.[depeq ] . if we write eq . [ depeq ] as @xmath19 , then we can solve for @xmath20 and @xmath14 using the 15 and 22 ghz data at @xmath21-arcsec resolution , obtaining ( for the e hotspot ) @xmath22 hz@xmath23 , @xmath24 per cent : we would then expect the degree of polarization at 8 ghz to be essentially zero ( about @xmath25 per cent ) whereas in fact there is a detection of polarization in this region at the @xmath26 level ( table [ fltab ] ) . the intrinsic polarization @xmath14 obtained here is also somewhat low , which suggests that the model is not accurate even for the two higher frequencies . however , the value of @xmath20 , which is the critical quantity for estimating physical parameters , simply depends logarithmically on the ratio of the polarizations at the two frequencies , and so is unlikely to be grossly wrong . applying the model , we can see from the _ hst _ data , assuming a cylindrical symmetry , that @xmath27 kpc : this would imply @xmath28 m@xmath29 nt kpc@xmath2 . if we further assume a simple model in which the differences are due to varying orientations of the magnetic field in an approximately uniform - density ( though possibly low - filling - factor ) medium , then , dropping numerical factors of order unity , @xmath30 m@xmath29 nt kpc@xmath2 . ( if there are in fact variations in @xmath31 , then the @xmath31 value derived from this equation is a weighted average . ) unfortunately , none of these three parameters , @xmath31 , @xmath32 and @xmath10 , is well constrained by existing observations . @xmath32 values of the order of 1 nt are energetically plausible from equipartition arguments , but this estimate is good to an order of magnitude at best . we know that @xmath33 kpc for the model above to be either necessary or applicable ; the fact that the degree of polarization of the hotspots continues to increase at higher radio resolutions ( table [ fltab ] ) suggests , however , that at least some structure is present on scales @xmath34 pc . the density is the most problematic quantity . clark et al . ( 1998 ) suggested that the density in the emission - line clouds in the nuclear regions is @xmath35 cm@xmath29 , but were unable to make measurements near the hotspots . electron densities @xmath34 cm@xmath29 are often quoted for eelr . however , it is important to emphasise that it is probably _ not _ the line - emitting material itself that is depolarizing the radio source , if it has a density close to this value . baum & heckman ( 1989 ) estimate constraints on the density and mass of the extended line - emitting material in 3c171 , based on the observed h@xmath37 luminosity and assuming case b recombination : for a filling factor of unity , they estimate @xmath38 cm@xmath29 . to obtain a density @xmath34 cm@xmath29 , a filling factor of @xmath39 would be required : in other words , if the line - emitting material has this density , it will have a very low covering fraction , and can not be responsible for depolarizing the radio source . for filling factors close to unity , a single - phase medium with a density close to the value given by baum & heckman can still just about explain the observations , requiring @xmath40 kpc for @xmath41 nt . however , if the eelr consists of dense ( @xmath34 cm@xmath29 ) , warm line - emitting clouds with a low filling factor , they can not be responsible for the observed depolarization . it must instead be due to a second phase of the ism . a plausible candidate is a second , hotter , more diffuse ionized phase , which would be expected to be present in shock - ionization models for the eelr : if the two phases are in approximate pressure equilibrium , then the density of the hot phase can be much lower than in the eelr clouds . if we take @xmath42 pc and @xmath43 nt , then the depolarization implies @xmath44 cm@xmath29 , which is a plausible density for the hot phase in the above scenario . for pressure balance with the colder , denser eelr clouds , that density would imply @xmath45 k : of course , we have no particular _ a priori _ reason to expect exact pressure balance in this dynamic situation . the non - detection of 3c171 in a deep _ rosat _ x - ray observation ( hardcastle & worrall 1999 ) sets quite strict limits on the properties of any hot gas in the system . i derive a @xmath46 upper limit on the absorbed _ rosat_-band ( 0.12.4-kev ) flux of the region of interest of @xmath47 ergs @xmath48 s@xmath2 . if we again assume cylindrical symmetry , with @xmath49 kpc and @xmath50 kpc , and consider a uniform medium , then fig.[nallow ] shows the constraints on gas density and temperature imposed by the x - ray limit together with the detection of depolarization . it can be seen that plasma of the temperature typically detected in the atmospheres of elliptical galaxies ( @xmath51 k ) can not be responsible for the depolarization : to avoid detection in the x - ray , its density would have to be low ( @xmath52 @xmath53 cm@xmath29 ) , implying scales @xmath54 kpc , which would be resolved by the radio data . this conclusion is valid so long as @xmath32 is not very much larger than 1 nt . the allowed region of parameter space in fig . [ nallow ] is more or less consistent , given the large uncertainties in all quantities involved ( the calculation is particularly uncertain at low temperatures ) , with the order of magnitude estimate of the temperature of the ` hot ' depolarizing medium given above : the x - ray data favour somewhat lower temperatures if @xmath55 cm@xmath29 , @xmath56 k. in the alternative scenario , involving a low - density eelr region , we expect @xmath57 k , which of course is entirely consistent with the x - ray upper limit . gas at temperatures around @xmath58 k can also in principle be detected in the optical via coronal emission lines ( e.g. graney & sarazin 1990 ) . if we adopt @xmath59 cm@xmath29 , then we can estimate upper limits on the temperature of the gas given the non - detection of any coronal lines in the spectra of clark et al .. graney & sarazin s results suggest that the most significant line for a plasma in ionization equilibrium ( not necessarily a good model for the material we are considering ! ) with @xmath60 k is the [ fe x ] 637.4 nm line . we derive an upper limit on the flux of this line from clark et al.s detection of [ o i ] at 636.3 nm , giving @xmath61 } \la 4 \times 10^{-16}$ ] ergs @xmath48 s@xmath2 . for @xmath59 cm@xmath29 , given our adopted geometry and the size of the extraction region of clark et al . we require the emissivity coefficient @xmath62 } \la 3 \times 10^{-27}$ ] ergs @xmath63 s@xmath2 . from figure 1 of graney & sarazin , this corresponds to @xmath64 k ; higher temperatures in the @xmath65@xmath66 k region would have given rise to detectable [ fe x ] emission in the optical spectra . thus the optical data , like the x - ray data , favour temperatures somewhat lower than @xmath58 k for the hot phase . again , this conclusion is sensitive to the choices of @xmath10 and @xmath32 , which determine the adopted value of @xmath31 . values of @xmath32 larger than 1 nt allow lower @xmath31 values and a hotter plasma . in addition , it is sensitive to the metallicity and ionization mechanism of the hot phase , which are unknown ; the key point here is simply to show that the presence of this phase is not ruled out by existing optical observations . to summarize , we are forced to one of two conclusions : either the eelr regions are very much less dense than the ` typical ' numbers quoted for @xmath67-k emission - line gas , and are distributed not in clumps but in a uniform screen across the source ; or , more probably , a second , less dense phase of the ism , with @xmath68 cm@xmath29 and a temperature likely to be around a few times @xmath69 k , is responsible for the depolarization . this second phase is plausibly shocked warm diffuse gas in the galaxy . in either case , the mass of the depolarizing medium , assuming it is cospatial with the eelr , would be @xmath70 m@xmath71 . the national radio astronomy observatory is a facility of the national science foundation operated under cooperative agreement by associated universities , inc . i am grateful to michael rupen for help with the special observing modes of the vla used in this project , and the vla analysts for help in implementing them in an observe file . i thank malcolm bremer for useful discussions about the nature of eelr , and clive tadhunter for supplying me with the _ hst _ image shown in section 1 . baum , s.a . , heckman , t. , 1989 , apj , 336 , 681 blundell , k.m . , 1996 , mnras , 283 , 538 burn , b.j . , 1966 , mnras , 133 , 67 clark , n.e . , axon , d.j . , tadhunter , c.n . , robinson , a. , obrien , p. , 1998 , apj , 494 , 546 de koff s. , baum , s.a . , sparks , w.b . , biretta , j. , golombek , d. , machetto , f. , mccarthy , p. , miley , g.k . , 1996 , apjs , 107 , 621 graney , c.m . , sarazin , c.l . , 1990 , apj , 364 , 561 hardcastle , m.j . , alexander , p. , pooley , g.g . , riley , j.m . , 1997 , mnras , 288 , 859 hardcastle , m.j . , worrall , d.m . , 1999 , mnras , 309 , 969 heckman , t.m . , van breugel , w.j.m . , miley , g.k . , 1984 , apj , 286 , 509 laing , r.a . , 1984 , in bridle a.h . , eilek j.a . , eds , physics of energy transport in radio galaxies , nrao workshop no . 9 , nrao , green bank , west virginia , p. 90 mccarthy , p.j . , 1993 , ara&a , 31 , 639 mccarthy , p.j . , baum , s. , spinrad , h. , 1996 , apjs , 106 , 281 mccarthy , p.j . , van breugel , w. , spinrad , h. , djorgovski , s. , 1987 , apj , 321 , l29 perley , r.a . , carilli , c.l . , 1996 , in carilli c.l . , harris d.e . , eds , cygnus a study of a radio galaxy , cambridge university press , cambridge , p. 168 tadhunter , c.n . , morganti , r. , robinson , a. , dickson , r. , villar - martin , m. , fosbury , r.a.e . , 1998 , mnras , 298 , 1035	i present high - frequency polarimetric radio observations of the radio galaxy 3c171 , in which depolarization is associated with an extended emission - line region . the radio hotspots , known to be depolarized at low radio frequencies , become significantly polarized above ( observers - frame ) frequencies of 15 ghz . there is some evidence that some of the faraday rotation structure associated with the emission - line regions is being resolved at the highest resolutions ( 0.1 arcsec , or 400 pc ) ; however , the majority remains unresolved . using the new radio data and a simple model for the nature of the depolarizing screen , it is possible to place some constraints on the nature of the medium responsible for the depolarization . i argue that it is most likely that the depolarization is due not to the emission - line material itself but to a second , less dense , hot phase of the shocked ism , and derive some limits on its density and temperature . galaxies : active galaxies : individual : 3c171 galaxies : ism
the inhomogeneous hubbard model we have studied is described by the hamiltonian [ eq : ihm ] h=-_i , j , t_ij ( c^_i , c_j,+ h.c.)+ _ iu_in_i , n_i , , where @xmath37 indicates nearest - neighbor sites , @xmath38 creates an electron on site @xmath10 with spin polarization @xmath39 , and @xmath40 . the usual ( homogeneous ) limit of this model is obtained by taking @xmath41 and @xmath42 . although we have studied a wider variety of patterns , to be concrete we will primarily focus on two inhomogeneous patterns of the hopping amplitudes : the checkerboard lattice and the striped lattice shown in fig . [ fig : lattice ] . unless otherwise stated , our discussion will focus on the case in which pbcs are applied in both the @xmath4 and @xmath43 directions . however , finite size effects will be estimated by comparing these results to those with `` twisted '' periodic boundary conditions . specifically , for a 4 @xmath23 4 system , pbc means identifying the sites @xmath44 and @xmath45 , while twisted pbc in the y direction means identifying @xmath44 and @xmath46 . we have also obtained results with open boundary conditions , but because of the large surface to volume ratio of the small systems studied , the proper interpretation of these results is unclear , and so we do not report them , here . the exact diagonalization is performed using the lanczos method @xcite , which has been employed by other people to extract some ground state properties of a @xmath0 ( homogeneous ) hubbard model @xcite . by taking advantage of conservation of the @xmath47 component of the spin and the @xmath48 symmetry of the checkerboard and @xmath49 symmetry of the striped model , we successfully reduce the dimension of the hilbert space from around @xmath50 to @xmath51 states and the ground state energies we found for @xmath52 agree with those of g. fano _ et al . _ @xcite . a complete group theoretical analysis which facilitates efficient implementation of the lanczos algorithm can be found in ref . . we present representative results on the _ pair - binding energy _ and the _ pair - field correlations_. we have much more extensive tables of results for various values of @xmath53 , @xmath54 and @xmath4 , and for various choices of boundary conditions . these additional results are available from the authors upon request . to better understand the pairing phenomena arising from repulsive interactions , we define the pair - binding energy : [ eq : pbenergy ] _ pb(x)=2e_0(m ) - [ e_0(m+1)+e_0(m-1 ) ] , where @xmath55 is the number of sites in the system , @xmath56 is the ground state energy with @xmath57 electrons ( _ i.e. _ @xmath58 holes doped into a `` neutral '' half - filled lattice ) , and @xmath59 is the `` concentration of doped holes . '' we will focus on the case in which @xmath58 is odd . thus , a positive pair - binding energy means that , given two isolated clusters with a mean doped hole density @xmath4 , it is energetically preferable to `` pair '' the doped holes so that one cluster has @xmath60 and the other cluster has @xmath61 doped holes . for a superconducting system , @xmath62 in the limit @xmath63 , where @xmath64 is the minimum value of the superconducting gap . while gaps can occur for other reasons ( _ e.g. _ cdw formation ) , a superconducting state , as far as we know , is the unique state that generically produces a gap for a non - zero range of @xmath4 in more than 1d . in a superconducting state with gapless ( nodal ) excitations , @xmath65 vanishes as @xmath63 , but only relatively slowly , in proportion to @xmath66 in 2d , so @xmath65 computed on finite size systems remains a good diagnostic of superconductivity . under generic circumstances in non - superconducting systems , the repulsive interactions between quasiparticles implies that @xmath65 is negative and @xmath67 as @xmath68 . a positive pair - binding energy on a small system could also indicate a tendency for phase separation . unambiguously distinguishing gap formation from phase separation can only be done by appropriate finite size scaling@xcite , which is beyond the reach of the present calculations . however , a gross test for phase separation is possible by testing whether further agglomeration of doped holes is favored . specifically , for @xmath69 even , we define @xmath70/2 .\ ] ] this is a crude approximation of the compressibility@xcite , which is negative in a system with a sufficiently strong tendency to phase separation . for the interesting range of @xmath2 and @xmath3 , we never find a negative value of @xmath71 and hence our system is unlikely to be phase separated . we have computed the pair - binding energy @xmath65 as a function of @xmath2 , @xmath3 , and @xmath4 on both the checkerboard and striped lattices with pbcs and twisted pbcs . contour plots of @xmath72 and @xmath73 for the checkerboard lattice are shown in respectively , in figs . [ fig : contour_pbe ] ( a ) and ( b ) , respectively . the global maximum can clearly be seen for @xmath7 and @xmath74 . @xmath72 is shown at fixed @xmath7 as a function of @xmath26 in fig . [ fig : pbenergy ] with both pbc ( squares ) and twisted pbcs ( triangles ) . a remarkable degree of insensitivity to boundary conditions is apparent for @xmath75 . on the coarse scale of @xmath26 in the main figure , it appears that @xmath65 is positive at small @xmath54 for all @xmath3 . however , this hides a subtlety at small @xmath54 , as shown in the inset to the figure in which the regime of small @xmath26 is shown on an expanded scale . specifically , as shown in fig . [ fig : cell_pbe ] , the pair - binding energy on an isolated square changes from positive to negative at @xmath76 . moreover , as discussed in ref . , it follows that the pair binding is positive with a non - zero limit as @xmath77 for @xmath78 , while @xmath79 ^ 2/t)$ ] is negative and vanishes as @xmath77 for @xmath80 . in fig . [ fig : doping_pbe ] , @xmath81 for the checkerboard model is shown at fixed @xmath82 as a function of @xmath54 for different values of @xmath83 , 3/16 , and 5/16 . a contour plot of @xmath72 for the striped lattice is shown in in fig . [ fig : contour_pbe ] ( c ) , and @xmath72 for fixed @xmath7 is shown as a function of @xmath54 for pbc and twisted pbc in fig . [ fig : pbenergy ] ( b ) . here , the results are apparently more sensitive to boundary conditions , so inferences concerning the thermodynamic limit are more difficult to reach . however , here , too , a global maximum of @xmath65 is reached for @xmath84 and @xmath85 . in the limit of vanishing @xmath26 , as follows from the results for an isolated ladder , fig . [ fig : cell_pbe ] ( dashed curve ) , @xmath65 is positive and non - vanishing for @xmath86 , negative and order @xmath87 for @xmath88 , again positive and non - vanishing for @xmath89 , and finally negative and order @xmath87 for @xmath90 . , as a function of @xmath26 on the checkerboard lattice.,width=294 ] , as a function of @xmath3 for an isolated @xmath91 plaquette ( solid curve ) and a @xmath92 ladder ( dashed curve ) . note that in the @xmath93 limit one can show @xmath94 for both clusters@xcite.,width=272 ] we have also studied the equal - time pair - field pair - field correlation function defined as d(,)=^_ij_kl , [ dij ] where the pair - field is ^_ij=(c^_i c^_j+c^_jc^_i ) . [ deltaij ] @xmath95 represents the bond between a pair of nearest - neighbor sites , @xmath96 , on the lattice . we focus on this correlation function for the largest possible separations , given the small system size : between a pair of parallel strong bonds separated by a distance 2 , _ i.e. _ bonds @xmath20 and @xmath22 in fig . [ fig : lattice ] , and a pair of perpendicular strong bonds separated by a distance @xmath97 , _ i.e. _ bonds @xmath20 and @xmath21 in fig . [ fig : lattice ] . we have computed these pairing correlations for both the checkerboard and striped models with @xmath98 , 2 , and 4 , _ i.e. _ for doped hole concentration @xmath99 , 1/8 , and 1/4 , and for a range of @xmath53 and @xmath54 . in fig . [ fig : paircorr](a ) , we show the pair correlation function for the checkerboard lattice at fixed @xmath82 as a function of @xmath54 for @xmath100 ( triangles ) and 1/4 ( squares ) . @xmath101 is represented by the dashed lines in the figure , and @xmath102 by the solid lines . the same quantities are shown for the striped lattice in fig . [ fig : paircorr](b ) . qualitatively similar results for both lattices have been obtained for @xmath99 , but the mangnitude of @xmath103 is an order of magnitude smaller than for @xmath104 , consistent with the expected mott insulating character of the undoped system . the positive sign of @xmath102 and the negative sign of @xmath101 are consistent with the d - wave character of the pairing ; in the thermodynamic , at large spatial separations , the sign of @xmath103 is determined entirely by the symmetry of the order parameter . for the checkerboard model , it is apparent that the magnitude of the pairing correlations are relatively weak both in the limit of strong inhomogeneity ( @xmath105 ) and of vanishing inhomogeneity ( @xmath106 ) . however , the strength of these correlations is only weakly dependent on @xmath54 for a broad range of intermediate values . moreover , in this intermediate regime , the strength of the pairing is quite similar for @xmath100 and @xmath107 . , as a function of @xmath26 on the checkerboard lattice . the subscript @xmath108 refers to the contribution from the strong and weak bonds , respectively . the open symbols indicate the data obtained at @xmath109 . as in eq . ( [ sma ] ) , @xmath110 times this quantity is the single - mode - approximation to the superconducting susceptibility.,width=294 ] it is more difficult to make a clear qualitative statement about the behavior of the striped model . one seemingly puzzling feature of the striped results deserves comment . in the limit @xmath111 , where the two ladders decouple , one might expect the correlation of the inter - ladder pair fields [ @xmath101 ] to vanish . however , since for @xmath113 , the pair binding energy on a @xmath92 ladder ( cube ) is positive , as shown in fig . [ fig : cell_pbe ] ( dashed curve ) , for @xmath114 and @xmath115 , there are two degenerate ground states , with the hole - pair on one or the other disconnected ladder . thus , even as @xmath26 tends to 0 , the ground - state wave function is a coherent superposition of these two states , and hence has substantial pair - field correlations . were one to study the finite temperature properties of this system , the coherence would be lost above a relatively small temperature @xmath116 . this illustrates the dangers of uncritically accepting evidence of strong superconducting correlations from the pair - field correlation function . we have also studied the expectation value of the pair - field operator between ground - states with @xmath69 and @xmath117 doped holes : _ ij m;0 \| _ ij \| m-2;0 [ avdelta ] where @xmath118 is the ground state with @xmath58 doped holes . this is not a gauge invariant quantity , in that there is an arbitrary choice of an overall phase ; by choosing the wave functions real , this is reduced to an overall sign ambiguity . however , the internal symmetry of the pairing state is manifest in this quantity . for the checkerboard case , for all @xmath3 and @xmath119 , the ground state symmetry as a function of @xmath58 alternates , @xmath120 for @xmath98 , @xmath121 for @xmath115 , @xmath120 for @xmath122 . thus , the pair - field operator that connects any two of these states must have precisely @xmath121 ( _ i.e. _ @xmath36 ) symmetry@xcite . in fig . [ fig : cc_sq ] , for fixed @xmath7 , we plot , as a function of @xmath54 , the gauge - invariant quantities , @xmath123 ( squares ) and @xmath124 ( triangles ) , where @xmath125 and @xmath126 are , respectively , the expectation value of @xmath127 for strong bonds @xmath17 ( within a square ) and weak bonds ( connecting two squares ) [ see eq . again , there is a clear indication that superconducting correlations are strongest for @xmath34 . two peculiarities are worth mentioning here : firstly , it is puzzling that the difference in the pair amplitudes on the weak and strong bonds is so large , @xmath128 , as clearly shown in fig . [ fig : cc_sq ] . this is a qualitative point that warrants further study . secondly , even in the limit @xmath129 , @xmath130 this is a consequence of the fact the ground - state is three - fold degenerate@xcite when @xmath131 and @xmath115 ; this degeneracy is lifted whenever @xmath132 . the @xmath133 superconducting susceptibility is expressible as _ a ( ) = _ where @xmath134 is the energy of the @xmath135th excited state with @xmath58 doped holes , and where we take ^(a ) 1 z_a _ f^(a)_ij_ij ; z_a= _ \|f^(a)_ij\|^2 . here , the chemical potential is appropriate to the case in which the ground - state has @xmath58 doped holes , @xmath136 @xmath137 > 2\mu > [ e_0(m ) - e_0(m+2)]$ ] . because of the d - wave symmetry , we always take @xmath138 to be positive on vertical bonds and negative on horizontal bonds . we consider three possible susceptibilities : @xmath139 in which where , @xmath140 on the strong bonds and 0 on the weak , @xmath141 in which @xmath140 on the weak bonds and 0 on the strong , and @xmath142 in which @xmath140 on all nearest - neighbor bonds . for simplicity , we have defined the susceptibility with respect to adding two - holes . another ( more conventional ) definition includes , as well , terms which remove two holes . .,width=294 ] as mentioned before , computing all the excited states that enter this sum is not feasible , but a lower bound estimate can be readily obtained in the `` single mode approximation '' , by approximating the sum by the single term involving the ground state with @xmath143 electrons . we refer to this as @xmath144 . in terms of the pair - field expectation values discussed in the previous section , & & _ s^(sma)=\|_s\|^2/ + & & _ w^(sma)=\|_w\|^2/ , and + & & _ t^(sma)=\|_s + _ w\|^2/2 . [ sma ] where @xmath145.$ ] thus , the quantities plotted in fig . [ fig : cc_sq ] can be viewed ( up to a factor of @xmath110 ) as the sma to the susceptibility at constant @xmath146 . moreover , it is worth mentioning that they are almost exact to the lower bound susceptibility as long as @xmath147 by tuning chemical potential properly . we now discuss the implications of our results . in particular , we are interested in using the present information to infer , as much as possible , the phase diagram of the checkerboard and striped hubbard models in the thermodynamic limit , @xmath63 . the results of our analysis and some additional speculations lead us to propose the qualitative phase diagram shown in fig . [ fig : pd_cb ] . there are two limits in which the checkerboard model simplifies : @xmath148 , where it can be studied using conventional diagramatic methods , and @xmath149 , where it reduces to weakly coupled squares which can be treated@xcite using degenerate perturbation theory in @xmath54 . this allows us to deduce the solid portions shown in the phase diagram in fig . [ fig : pd_cb ] without resorting to the present numerical results . for @xmath150 , and in the thermodynamic limit , the non - interacting fermi - surface depends qualitatively on @xmath54 . there are four bands since each unit cell contains four sites and @xmath151 and @xmath152 are the basis vectors of the reciprocal lattice . for @xmath153 , there are two electron pockets enclosing the @xmath58 points @xmath154 and @xmath155 respectively , and one hole pocket , enclosing the `` nodal point '' @xmath156 , as shown in fig . [ fig : fermisurface](a ) . the hole pocket shrinks to a point as @xmath26 approaches the critical value for a lifshitz transition , @xmath157 . for @xmath158 , only the two electron pockets enclosing the @xmath58 points remain , as shown in fig . [ fig : fermisurface](b ) . for @xmath159 , and 1/4 , @xmath160 is , respectively , @xmath161 , and @xmath162 . this lifshitz transition appears in the conjectural phase diagram in fig . [ fig : pd_cb ] as the boundary between two fermi liquid phases - fl1 ( with two electron pockets plus one hole pocket ) and fl2 ( with only two electron pockets ) . in sketching this figure , we have assumed that the fermi liquid phases are stable in the presence of a small repulsive @xmath3 ; it is likely that this is not strictly the case , since there is probably a kohn - luttinger instability of any fermi liquid @xcite , but if this occurs , it is on such a low energy and temperature scale that it can be neglected for present purposes . for ( a ) @xmath163 , showing the two electron and one hole fermi pockets , and ( b ) @xmath164 , with only the two electron pockets . , width=294 ] the small @xmath26 portion of the phase diagram in fig . [ fig : pd_cb ] was derived previously in refs . and . for @xmath165 , where an isolated hubbard square has a positive pair - binding energy , there exists a nodeless d@xmath166-wave superconducting phase . for @xmath167 , there is a third fermi liquid phase , fl3 , which has the same fermi surface topology as fl2 . finally , for @xmath168 ( where the isolated square with one doped hole has a fully polarized spin 3/2 ground state ) , the system exhibits an exotic , spin 3/2 fermi liquid phase , fl4 , if @xmath169 , while it phase separates into two insulating antiferromagnetic charge ordered phases if @xmath170 . the superconducting state at small @xmath26 arises from the geometry of the square and the strong correlations produced by @xmath3 . however , across the phase boundary between the superconducting phase and fl2 or fl3 , it is reasonable to view the onset of superconductivity as a bcs - like fermi surface instability . in this limit , the fact that the d - wave superconductor is nodeless can be understood as being a simple consequence of the fact that the fermi surface does not intersect the line of d - wave gap nodes , so there are no gapless quasiparticles . to obtain a more complete phase diagram , we have used the insights obtained from the present exact diagonalization studies . it is , of course , not clear to what extent the results from small system studies can be extrapolated to the @xmath63 limit . however , at least in the case of the pair - binding energy , since the results appear to be relatively insensitive to boundary conditions for @xmath171 , we feel that we can use these results as the basis of a set of plausible conjectures . these are shown as the dashed lines in fig . [ fig : pd_cb ] . notice that we have superposed the conjectured phase diagram on the contour plot of the pair - binding energy from fig . [ fig : contour_pbe](a ) . where the pair - binding energy on the 4 @xmath23 4 system is large and insensitive to boundary conditions , we feel that we are on sound grounds when we speculate that this corresponds to a well developed , nodeless d - wave superconducting state in the @xmath68 limit . unfortunately , where @xmath65 is small and/or sensitive to boundary conditions , this could mean that in the @xmath63 limit , the system has entered a gapless phase , _ i.e. _ a fl or a nodal d - wave superconductor . however , it could simply mean that the minimum gap of a nodeless d - wave superconductor is small and the corresponding correlation length is long . in drawing our conjectural phase diagram , we have assumed the former , and so , to the extent possible , we have drawn phase boundaries between the nodeless d - wave phase and various gapless phases along contours separating the region of `` large '' and boundary condition insensitive pair - binding energy , to regions with smaller , and/or strongly boundary condition dependent pair - binding . clearly , the upper portion of the phase diagram ( @xmath172 ) is the most speculative portion , including the entire region in which nodal d - wave superconductivity occurs . the existence of a tetracritical point in the conjectured phase diagram , with the additional implication that there exists a nodal d - wave superconducting state for large enough @xmath26 , follows from the nature of the known phases along the edges of the phase diagram . in particular , if there is a direct , continuous transition from fl1 to a d - wave superconductor , the d - wave superconductor must be nodal . however , we can not rule out the possibility of a direct first order transition from fl1 to a d - wave superconductor , in which case the tetracritical point could be replaced by a bicritical point , and the d - wave superconductor could always be nodeless . the portion of the phase diagram with @xmath173 could also exhibit additional broken symmetry phases @xcite . it is impossible , with any degree of confidence , to use the present results to infer anything new about the phase diagram of the striped hubbard model in the @xmath63 limit . for @xmath174 , the model becomes an array of weakly connected two - leg ladders , a problem which was studied previously by e. arrigoni _ _ in ref . from that work , we know that for @xmath175 , there exists a nodeless `` d - wave like '' superconducting state over a very broad range of @xmath53 . in contrast , the oscillatory behavior of the pair - binding energy seen in the fig . [ fig : contour_pbe](c ) as @xmath176 is a special feature of the a @xmath92 ladder , which does not occur in a system of weakly coupled , infinitely - long two - leg ladders . combining this observation with the strong boundary condition dependence of the pair - binding energy apparent in fig . [ fig : pbenergy](b ) ( even when the twist in the boundary conditions is applied in the direction perpendicular to the stripe direction ) , we are forced to conclude that the results on the 4 @xmath23 4 stripe lattice are not representative of the @xmath63 limit . we therefore do not venture to draw even a conjectural phase diagram for this system . nevertheless , on the basis of the fact that , in fig . [ fig : contour_pbe](c ) , there is an extended region with relatively large pair - binding energy when @xmath2 is a substantial fraction of @xmath1 , makes plausible the speculation made in ref . that the nodeless superconducting state grows in strength for a substantial range of non - infinitessimal @xmath54 . furthermore , if indeed there is a nodal d - wave state for @xmath54 near 1 , there must also exist a lifshitz - type phase transition to a superconducting state with gapless quasi - particles at a @xmath53 dependent critical value of the inter - ladder coupling @xcite . one issue that is often ignored in discussions of htc is the role of the longer - range coulomb interactions . d - wave pairing avoids the obvious deliterious effects of the on - site hubbard repulsion between electrons , but is known@xcite to be fairly sensitive to longer range repulsive interactions . to address this issue , we have computed the effect on the pair - binding energy of a nearest - neighbor repulsion , @xmath177 , between electrons . indeed , we always find that the pair - binding energy decreases , more or less linearly , with increasing @xmath177 . however , where the pair - binding is strong for @xmath178 , it remains positive up to rather large values of @xmath177 . for instance , for the checkerboard lattice under optimal conditions , @xmath7 and @xmath179 , @xmath65 is an essentially linear function of @xmath177 which vanishes at @xmath180 . in conclusion , we have studied inhomogeneous hubbard models , primarily with checkerboard and striped patterns , on a @xmath0 square lattice with periodic boundary conditions by exact diagonalization . although the existence of the htc in the uniform hubbard model is still a controversial issue @xcite , we have produced clear evidence that , without considering other non - electronic degrees of freedom such as phonons , strong pairing of electrons can be achieved from purely repulsive interactions if certain modulations of the electronic structure are introduced . non - monotonic dependence of the pair - binding energy and the pair - field pair - field correlators on the degree of inhomogeneity ( @xmath181 ) were found to be generic . this observation supports the notion that there is an optimal inhomogeneity for high temperature superconductivity @xcite . since exact diagonalizaiton studies can not access significantly larger systems , it has not been possible to carry out finite size scaling to corroborate this conclusion . however , we hope that the present results will stimulate further work on larger system using more efficient numerical tools such as density matrix renormalization group ( dmrg ) @xcite or quantum monte carlo ( qmc ) . we thank e. fradkin and d. j. scalapino for helpful discussions . this work was supported in part by doe grant no . de - fg02 - 06er46287 . is supported by a sgf at stanford university . computational resources were provided by slac at stanford university and some of the computations were performed at the cscs manno , switzerland . for instance , j. k. cullum and r. a. willoughby , _ lanczos algorithms for large symmetric eigenvalues computations _ ( birkhauser , boston , 1985 ) . e. dagotto , f. ortolani , and d. scalapino , phys . b * 46 * , 3183(r ) ( 1992 ) ; p. w. leung , z. liu , e. manousakis , and p. e. oppenheimer , _ ibid _ * 46 * , 11779 ( 1992 ) . g. fano , f. ortolani , and a. parola , phys . rev . b * 42 * , 6877(r ) ( 1990 ) . g. fano , f. ortolani , and f. semeria , int . b3 * , 1845 ( 1989 ) ; g. fano , f. ortolani , and a. parola , phys . rev . b * 46 * , 1048 ( 1992 ) . v. j. emery and s. a. kivelson , physica c * 209 * , 597 ( 1993 ) . s. chakravarty , l. chayes , and s. a. kivelson , lett . * 23 * , 265 ( 1991 ) . d. j. scalapino and s. a. trugman , philos . mag . b * 74 * , 607 ( 1996 ) . w .- f . tsai and s. a. kivelson , phys . b * 73 * , 214510 ( 2006 ) ; _ ibid _ * 76 * , 139902 ( 2007 ) . h. yao , w .- f . tsai , and s. a. kivelson , phys . rev . b * 76 * , 161104(r ) ( 2007 ) . w. kohn , and j. m. luttinger , phys . lett . * 15 * , 524 ( 1965 ) . e. arrigoni , e. fradkin , and s. a. kivelson , phys . b * 69 * , 214519 ( 2004 ) . g. hager , g. wellein , e. jeckelmann , and h. fehske , , 075108 ( 2005 ) e. altman and a. auerbach , phys . b * 65 * , 104508 ( 2002 ) . m. granath , v. oganesyan , s. kivelson , e. fradkin , and v. emery , phys . lett . * 87 * , 167011 ( 2001 ) . s. r. white , s. chakravarty , m. p. gelfand , and s. a. kivelson , , 5062 ( 1992 ) . for recent work , see t. aimi and m. imada , j. phys . jpn . * 76 * , 113708 ( 2007 ) , and references therein . s. chakravarty , and s. a. kivelson , phys . b * 64 * , 064511 ( 2001 ) . i. martin , d. podolsky , and s. a. kivelson , phys . b * 72 * , 060502(r ) ( 2005 ) . k. aryanpour , t. paiva , w. e. pickett , and r. t. scalettar , phys . b * 76 * , 184521 ( 2007 ) . y. l. loh and e. w. carlson , phys . rev . b * 75 * , 132506 ( 2007 ) . e. arrigoni and s. a. kivelson , phys . b * 68 * , 180503(r ) ( 2003 ) . s. r. white , private communication .	we report the results of exact diagonalization studies of hubbard models on a @xmath0 square lattice with periodic boundary conditions and various degrees and patterns of inhomogeneity , which are represented by inequivalent hopping integrals @xmath1 and @xmath2 . we focus primarily on two patterns , the checkerboard and the striped cases , for a large range of values of the on - site repulsion @xmath3 and doped hole concentration , @xmath4 . we present evidence that superconductivity is strongest for @xmath3 of order the bandwidth , and intermediate inhomogeneity , @xmath5 . the maximum value of the `` pair - binding energy '' we have found with purely repulsive interactions is @xmath6 for the checkerboard hubbard model with @xmath7 and @xmath8 . moreover , for near optimal values , our results are insensitive to changes in boundary conditions , suggesting that the correlation length is sufficiently short that finite size effects are already unimportant . the relatively large energy scales and short coherence lengths involved in high temperature superconductivity ( htc ) imply that theories of the `` mechanism '' must involve different considerations than the conventional bcs theory of low temperature superconductivity ( ltc ) . the theory of ltc can be treated in the context of fermi liquid theory , in which the strong effects of electron - electron repulsions are resolved at high energy , so that pairing is a fermi surface instability triggered by weak , retarded , induced attractive interactions . the theory of htc must treat the strong local repulsions between electrons directly , as fermi liquid theory is certainly not valid at short distances and high energies . conversely , since the physics of htc is relatively local , numerical studies of finite size systems can , plausibly , resolve questions concerning the mechanism so long as the system sizes are large compared to the coherence length , a condition that could not remotely be envisaged for ltc . in the present paper , we report exact diagonalization studies of the low energy states of the hubbard model on a square lattice ( fig . [ fig : lattice ] ) , with hopping matrix elements , @xmath9 , between pairs of sites , @xmath10 and @xmath11 , and with a short - range repulsive interaction , @xmath12 , between two electrons on the same site . we propose to address the following sharply posed question , related to the physics of the mechanism : what form of the hamiltonian ( _ i.e. _ values of @xmath13 and @xmath14 ) maximizes @xmath15 subject to the constraints that 1 ) @xmath9 is short - ranged and bounded , _ i.e. _ @xmath16 for all @xmath17 , and that 2 ) the interactions are repulsive , @xmath18 for all @xmath10 ? constraint 2 represents a theoretical prejudice that htc derives directly from the strong repulsive interactions between electrons , and could be relaxed in future studies . the requirement that @xmath9 is bounded avoids the trivial answer that , for any hamiltonian that supports superconductivity , @xmath15 can be doubled by simply doubling the hamiltonian . on the solid `` strong '' bonds and @xmath19 on the dotted `` weak '' bonds . the three labeled bonds , @xmath20 , @xmath21 , and @xmath22 are the focus of our study of pairing correlations.,width=249 ] the largest systems we can readily study are 4 @xmath23 4 . for such small systems , there is no direct way to extract a @xmath15 . we have thus introduced other benchmarks of the strength of the superconductivity that can be readily computed on finite size systems , especially the `` pair binding energy , '' defined in eq . ( [ fig : pbenergy ] ) , and the strength of the pair - field correlations defined in eqs . ( [ dij])-([avdelta ] ) . the full optimization problem we have proposed would be prohibitively time consuming . we have therefore concentrated on a restricted , highly symmetric subset of all possible hamiltonians , with particular focus on the symmetric checkerboard and stripe patterns shown in figs . [ fig : lattice](a ) and [ fig : lattice](b ) . we take the hopping matrix elements to be @xmath1 on the solid ( `` strong '' ) bonds in the figure , and @xmath19 on the dotted ( `` weak '' ) bonds . for @xmath24 , the system is the homogeneous hubbard model , while for @xmath25 , the system consists of disconnected hubbard squares or ladders . thus , as we vary @xmath26 from 0 to @xmath1 , we vary the `` degree of inhomogeneity '' . for instance , as shown in the contour plot in fig . [ fig : contour_pbe ] , the pair - binding energy of the checkerboard lattice with periodic boundary conditions ( pbc ) is largest for @xmath7 and @xmath27 . the concentration , @xmath4 , of `` doped holes '' per site ( _ i.e. _ the deviation from one electron per site ) can only take discrete values ; among those , the optimal pair - binding is largest when @xmath28 ( as in the contour plot in fig . [ fig : contour_pbe](a ) ) , somewhat smaller when @xmath29 ( fig . [ fig : contour_pbe](b ) ) , and is always negative ( pair - repulsion ) for @xmath30 . indeed , of all forms of the hamiltonian we have explored to date , the checkerboard hubbard model with these parameters has the largest pair - binding energy we have found . moreover , by changing boundary conditions to twisted pbc , shown as the solid ( red ) triangles in fig . [ fig : pbenergy](a ) , we see that at fixed @xmath7 , as a function of @xmath26 in the range @xmath31 , the results are sensitive to change of boundary conditions only at the 20% level . we therefore infer that , _ the checkerboard hubbard model in the thermodynamic limit , has a maximum value of the pair - binding @xmath32 for @xmath33 , @xmath34 , and @xmath35 . _ from analysis of the ground - state symmetry and of the pair - field correlations , we can identify the dominant superconducting correlations on this system as d - wave ( @xmath36 ) .
atomic - scale magnetic structures on surfaces @xcite are of current interest for several reasons . primarily , they display intriguing physical properties in their own right . magnetic atoms on surfaces , simple or complex , can display large magnetocrystalline anisotropy which differs from the bulk . the spin can be large , or quenched by electronic effects such as the kondo effect . and coupling between spins can be via direct overlap , rkky or superexchange . secondarily , these systems are compelling because of their parallels with other nano - scale systems quantum dots , magnetic multilayers , magnetic impurities in thin films , to name just a few . finally , there are the possible applications to nano - scale magnetic bits and future magnetic devices . a large net spin and magnetic anisotropy are required for atomic - scale magnetic structures to be used as practical nano - scale magnetic bits . a possible way to obtain a large magnetic moment is through a ferromagnetic coupling between transition metal atoms . however , ferromagnetic coupling is rare in transition metal complexes , @xcite that is , when the transition metal atom is bonded to a nonmetallic atom . we describe below our studies of such a system . in a scanning tunneling microscope ( stm ) measurement @xcite of a ti atom placed on a cu@xmath0n / cu(100 ) surface , it was found that the ti exhibits very different magnetic properties than in gas phase . in the following work , we use density functional theory ( dft ) to study the atomic spin of a single ti atom , and exchange coupling of a dimer of ti atoms , placed on a single layer of cu@xmath0n on a cu(100 ) surface . the cu@xmath0n layer is used as an insulating layer to isolate the spin of a magnetic adatom from the metal electrons of the cu(100 ) surface @xcite . hereafter , the cu@xmath0n / cu(100 ) surface will be referred to as the cun surface . we study exchange coupling between ti atoms in two different environments : ( i ) a square lattice of ti on the cun surface and ( ii ) a dimer of ti atoms deposited on the cun surface . we use spin - polarized dft , as implemented in quantum - espresso @xcite , within a pseudopotential formalism using a plane wave basis with a cut - off of 30 ry . a higher cut - off of 240 ry was used for the augmentation charges introduced by the ultrasoft pseudopotential @xcite . we use the generalized gradient approximation ( gga ) for the exchange correlation interaction with the functional proposed by perdew , burke and ernzerhof . @xcite an on - site coulomb interaction ( u ) for ti was employed , with @xmath1 taken to be 4.7 ev as calculated using a constraint - gga method @xcite . to improve the convergence , a gaussian smearing of width 0.01 ry was adopted . brillouin zone integrations for the ( @xmath2 ) surface cell of cu(100 ) were carried out using a ( @xmath3 ) mesh of @xmath4-points . we obtained the bulk lattice parameter for cu as 3.67 , which compares well with the experimental value of 3.61 @xcite . to simulate the cun surface , we use a symmetric slab of three to five atomic layers of cu , with a layer of cun above and below . periodic images of the slab were separated by a vacuum of 15 along the z @xmath5 $ ] direction . the top view of the cun surface is shown in fig . [ cun - strct - charge]a . the unit cell of the cun surface ( shown by the black square ) consists of two cu atoms and one n atom . we find that the n atoms are 0.18 above the top cu atoms in a fully relaxed structure . the distance between the first and second layers of cu is 1.97 which compares well with the all - electron result of 1.91 @xcite . we find that each cu atom in the surface unit cell loses 0.7 electrons to the n ( presumably due to the electronegative nature of n ) . thus , cu and n atoms form a square network of dipoles on the surface , rendering a ( nominally ) insulating character to the surface . as shown in fig.[cun - strct - charge]a , along the x - axis , two surface cu atoms are separated by a hollow site and along the y - axis , they are separated by a n - atom . these directions will be referred as hollow - axis and n - axis , respectively . the charge density of the cun surface along the n- and hollow - axis is shown in fig.[cun - strct - charge](b ) and ( c ) respectively . notice that along the n - axis , charge gets accumulated in the top layer . ) unit cell of the cun surface . along the x - axis , surface cu atoms are separated by a hollow site ( it is marked by ` h ' ) and along the y - axis , they are separated by a n atom . side views of charge density produced by a cut along the ( b ) n - axis and ( c ) hollow - axis.,width=317 ] to calculate the spin of ti on the cun surface , a ti atom is deposited on top of a cu atom , following the stm experiments @xcite . at one monolayer coverage of ti , the ( @xmath6 ) surface unit cell consists of one n atom , one ti atom and two cu atoms as shown in fig.[cun - strct - charge]a . a constrained - gga @xcite calculation @xcite yields @xmath7 ev for ti in this configuration . however , the value of @xmath1 for ti would presumably be different at lower coverages of ti . in order to understand the effect of u on the spin of ti , we do calculations for a range of values of @xmath1 for the ( @xmath6 ) unit cell . the results for distances between ti and its nearest neighbor atoms in the surface , angle subtended by ti - n - ti , and the spin of ti are shown in table [ structure ] . we find that ti - n and ti - cu distances increase as @xmath1 increases , with the net effect of a rising ti and decreasing ( becoming sharper ) ti - n - ti angle . most importantly , as u increases the spin of ti approaches that of the gas phase value of 1 . as an added complexity , we find that the initial magnetization of ti affects the final calculated ground state , indicating a complex energy minimization landscape . hence we try several initial magnetizations and take the final state corresponding to the lowest energy . for u=4.7 , we show the data corresponding to two such optimized structures ( s - i and s - ii ) obtained by varying the initial magnetization ( see table [ structure ] ) . the optimized structure corresponding to s - i is the lowest energy structure ( s - i is lower in energy than s - ii by 0.3 ev ) , showing the spin of ti to be 0.75 , indicating possible mixed valent behavior for a monolayer of ti . [ cols="^,^,^,^,^,^,^ " , ] we have drawn conclusions so far about ti - ti coupling based on the calculations in lattices , where the situation is more complicated because one not only has the nearest neighbor ( nn ) interactions but also has next nn ( nnn ) interactions and so on . to simulate a ti - ti dimer on the surface we use a larger unit cell of ( @xmath8 ) with two and three lattice units along the hollow- and n - axis respectively . interestingly , we find ferromagnetic coupling along the n - axis and antiferromagnetic along the hollow - axis , the same ground states as for the case of ( @xmath6 ) lattices . along the n - axis , the energy difference ( @xmath9 ) is -16.5 mev compared to -16.1 mev for the ( @xmath6 ) case . along the hollow - axis the energy difference is 143.9 mev compared to 106.8 mev for the ( @xmath6 ) case . thus , a ti lattice and a dimer show a similar trend and strength of coupling . it confirms our assumption of primarily nearest neighbor interactions in a ti lattice on the cun surface . notice that the distance between ti and the cu atom below it is 2.56 for both the ( @xmath6 ) case , and the ( @xmath8 ) case for coupling along the hollow - axis . however , when the dimer is placed along the n - axis the distance between ti and the cu below it increases slightly to 2.65 . the ti - n - ti angle is 135.6 degrees for the ( @xmath6 ) case which is close to 142.9 degrees for the ( @xmath8 ) case . in fig . [ spin - dimer ] , we plot the spin density for the ti dimer along the n - axis ( fig.[spin - dimer]a ) and the hollow - axis ( fig.[spin - dimer]b ) . a significant amount of induced spin - polarization around the n atom can be seen from the figure . ferromagnetic coupling between ti atoms along the n - axis is established by having an opposite spin n atom both between the ti atoms and at opposite ends . for anti - aligned spin configuration along the n - axis , the n atom becomes a single - atom antiferromagnet with a net spin of zero . along the hollow - axis , when spin on both the ti atoms is aligned , a dramatic anisotropy in the spin polarization of the ti develops , with a direct overlap established over the hollow site ( fig.[spin - dimer]b ) . the stretching of the ti bonds in this case case is striking , and suggestive that higher symmetry considerations may be coming into play . however , when spins are antialigned , no such elongation of spin polarization occurs . in both the cases , n atoms on the sides of the two ti atoms develop a spin polarization opposite to that of the ti . the primary sources of exchange coupling between the ti atoms are superexchange @xcite , rkky @xcite , and direct overlap / direct exchange @xcite . the coupling between the adatoms can be direct , if the wave functions should overlap , or rkky , if the influence of the cu in the layers below is strong enough . along the n - axis , the center n atom becomes a natural source for a superexchange coupling between ti atoms , ruling out rkky which would need to take an indirect route under the n atom , a much longer route than directly across the n for superexchange . along the hollow - axis , however , there is no convenient single atom to hop across for superexchange , rather the sea of conduction electrons from the underlying and intervening cu . ( unless one is to consider superexchange via the second - layer cu , an unlikely candidate . ) in this case rkky and direct overlap become more likely . indeed for an aligned spin configuration , we observe a direct overlap forming , as discussed above . however , the lowest energy state for coupling along the hollow - axis is antiferromagnetic , and we conclude that in this case it is likely due to rkky coupling . this could be tested experimentally by varying the ti - ti distance and measuring the exchange coupling ; however , only certain discrete lattice positions would be possible . ) plot for ferromagnetic ( top pannel : ( a ) and ( b ) ) and antiferromagnetic ( bottom pannel : ( c ) and ( d ) ) configuration of a ti dimer along the ( a ) n - axis and ( b ) hollow - axis . the most energetically favorable configurations are ferromagnetic across a n ( top left ) and antiferromagnetic across a void ( bottom right).,width=226 ] we find that a ti atom has spin-1 on the cun surface instead of spin-1/2 as found in the experiments @xcite . as a possible explanation , we propose that there should be a two - stage kondo effect for this system . at high temperatures , ti starts with spin-1 , which then undergoes a transition to spin-1/2 with a first - stage kondo effect . at a lower temperature the second stage kondo effect brings it down from spin-1/2 to a spin - zero , and it is this spin-1/2 kondo effect that is seen in the low - temperature stm . this prediction can be tested by measuring the local density of states at higher temperatures than 0.5 k. we find a ferromagnetic coupling along the n - axis and antiferromagnetic along the hollow - axis , for both the lattice and dimer of ti on the cun surface . ti lattice and dimer have a similar trend as well as strength of coupling . this indicates that interactions between ti atoms in the lattice configuration are local ; and a marked spin striped phase is found as the ground state of the lattice . we find a ferromagnetic coupling along the n - axis due to superexchange , with secondary contributions from direct exchange . we also postulate that the antiferromagnetic coupling along the hollow - axis is primarily due to rkky interactions , with a smaller direct exchange component .	we have studied the spin and exchange coupling of ti atoms on a cu@xmath0n / cu(100 ) surface using density functional theory . we find that individual ti have a spin of 1.0 ( i.e. , 2 bohr magneton ) on the cu@xmath0n / cu(100 ) surface instead of spin-1/2 as found by scanning tunneling microscope . we suggest an explanation for this difference , a two - stage kondo effect , which can be verified by experiments . by calculating the exchange coupling for ti dimers on the cu@xmath0n / cu(100 ) surface , we find that the exchange coupling across a ` void ' of 3.6 is antiferromagnetic , whereas indirect ( superexchange ) coupling through a n atom is ferromagnetic . for a square lattice of ti on cu@xmath0n / cu(100 ) , we find a novel spin striped phase .
generalization of the thermostatistics is known to be based on the deformation procedure of both logarithm and exponential @xcite . the simplest way to introduce these functions into the thermostatistics scheme is to consider the equation of motion for dimensionless volume @xmath6 of the supported phase space ( @xmath7 , @xmath8 being dirac - planck constant and particle number ) . in the course of evolution of the ensemble with statistical weight @xmath9 and entropy @xmath10 , the variation rate of the phase space volume is governed by the following equation @xcite @xmath11 from this , one follows the relation @xmath12 whose integration gives the entropy related to the whole statistical weight @xmath13 : @xmath14 here , we take into account that entropy of a single state @xmath15 vanishes . in the case of the smooth phase space , one has trivial relation @xmath16 whose insertion into the relation ( [ 2 ] ) derives to the boltzmann entropy @xmath17 . however , complex systems has fractal phase space with the dimension @xmath18 , so that relation between the statistical weight and the corresponding volume should be replaced by the power - law dependence @xmath19 where reduced fractal dimension @xmath20 is introduced . insertion of eq . ( [ 3 ] ) into integral ( [ 2 ] ) gives the expression @xmath21 which is reduced to the tsallis logarithm @xmath22 where non - additivity parameter @xmath23 is replaced by inverted value @xmath24 of the reduced fractal dimension @xmath0 of the phase space . naturally , this expression is reduced to the boltzmann entropy in the limit @xmath25 . adduced formalism is based on the proposition that the phase space is a monofractal determined by single dimension @xmath0 . the aim of this letter is to generalize tsallis thermostatistics onto the multifractal phase space with a spectrum @xmath2 . such a generalization for arbitrary distribution @xmath2 is carried out in section 2 . discussion in section 3 shows that overall representation of the thermostatistics based on the multifractal phase space demands of the consistent consideration of input , escort and ` physical ' distributions . an optimization of the spectrum @xmath2 is considered in section 4 . section 5 concludes our treatment . according to the usual recipe @xcite , the passage to multidractal phase space is provided with replacement of the power function ( [ 3 ] ) by the expression @xmath26 where @xmath23 is multifractal index , multiplier @xmath27 takes into account specific number of monofractals with dimension @xmath0 within the multifractal . as a result , the statistical weight , playing the role of the multifractal measure , takes the form @xmath28 where @xmath29 is the density distribution over dimensions @xmath0 . using the method of steepest descent , we arrive at the modified power law @xmath30 which generalizes the simplest relation ( [ 3 ] ) due to replacement of the bare fractal dimension @xmath0 by the multifractal function @xmath31 . the conditions of application of the steepest descent method @xmath32 allow us to find special fractal dimension @xmath33 related to given parameter @xmath23 . above consideration shows that the passage from monofractal phase space to multifractal one is obtained by replacement of the single dimension @xmath0 by the function @xmath34 that monotonically increases , taking value @xmath35 at @xmath36 and @xmath37 at @xmath38 . limit behaviour of the function @xmath34 is characterized by the asymptotics @xmath39 physical domain of the @xmath23 parameter variation is bounded by condition @xmath40 which ensures positive values of the function @xmath41 . as a result , we can use well - known tsallis formalism of the non - extensive statistics where role of the non - additivity parameter plays inverted value @xmath3 of the multifractal function @xmath31 . thus , the entropy in dependence of the probability distribution @xmath42 has the form @xcite @xmath43 with accounting the conditions @xmath44 this expression arrives at the generalized distribution over energy levels @xmath45 : @xmath46 where @xmath47 is the partition function and the deformed exponential is determined in the usual manner : @xmath48^{\frac{1}{{\bar\tau}-1}}\ { \rm at}\ 1+({\bar\tau}-1)x>0,\\ 0\ \ \qquad\quad\qquad\qquad\quad\qquad\quad{\rm otherwise}. \end{array } \right . \label{d}\end{aligned}\ ] ] the normalization condition ( [ 13 ] ) gives the relation @xmath49}-\left[{\bar\tau}(q)-1\right]\beta e_q . \label{15}\ ] ] in spite of the formal similarity of above expressions with the same of the tsallis statistics , a principle difference takes place : in the last case , the limit of extensive systems is reached , when non - additivity parameter takes value @xmath38 , whereas the multifractal phase space is reduced to monofractal one with dimension @xmath50 in the opposite limit @xmath51 . in this way , the step - function spectrum @xmath34 , being @xmath52 at @xmath40 and @xmath37 otherwise , relates to the smooth phase space . thermodynamic functions of the model under consideration can be found analogously to the tsallis non - extensive scheme @xcite . however , related expressions are very cumbersome even in the simplest case of the ideal gas @xcite and amount to the usual form only within the slightly non - extensive limit @xcite . at the same time , developed scheme allows one to use thermodynamic formalism of multifractal objects @xcite . within the latter , the role of a state parameter plays the multifractal index @xmath23 , whose variation may arrive at phase transitions if the dependence @xmath34 has some singularities . it is worthwhile to stress that developed scheme arrives directly at related singularities of thermodynamic functions type of the internal energy ( [ 18 ] ) ( see below ) , the entropy ( cf . eq.([4 ] ) ) @xmath53 and the free energy @xmath54 in accordance with the non - extensive thermostatistics @xcite the distribution ( [ 14 ] ) plays the role of the escort probability related to the input one @xmath55^{{\bar\tau}(q ) } , \label{26}\ ] ] being normalized by the condition @xmath56 easily to convince , the expression ( [ 26 ] ) can be reduced to pseudo - gibbs form @xmath57\label{28}\ ] ] if one introduces effective value of the inverted temperature @xmath58}\beta\label{29}\ ] ] and determines the free energy @xmath59 in the usual manner : @xmath60 it is characteristically , the escort distribution ( [ 14 ] ) , determined with the physical temperature @xmath61 , is normalized by the standard condition ( [ 13 ] ) , whereas the input probability ( [ 26 ] ) , being deformed exponential ( [ 28 ] ) , is determined with non - physical temperature ( [ 29 ] ) to derive to the deformed free energy ( [ 30 ] ) . thus , both input and escort distributions appear as complementary ones . in this way , the escort distribution ( [ 14 ] ) is deformed with the inverted index @xmath62 , whereas the input probability ( [ 28 ] ) with the index @xmath34 itself . this exhibits the known ` @xmath63'-duality of tsallis thermostatistics @xcite . according to @xcite the physical distribution is neither escort nor input probabilities , but the following one @xmath64 it corresponds to the condition @xmath65 instead of the second equation ( [ 13 ] ) . easily to show , the probabilities ( [ 26 ] ) , ( [ a ] ) are connected by the relation @xmath66}p(\varepsilon_i - e_q ) \label{cc}\ ] ] to derive the distribution @xmath67 . \label{c}\ ] ] in the case of continuous energy spectrum characterized with the density distribution @xmath68 , the internal energy related to the condition ( [ b ] ) takes the form @xmath69 extreme value of @xmath70 is reached at the condition @xmath71 where prime denotes differentiation over @xmath72 . usually , the density function is reduced to the power law @xmath73 , @xmath74 , so that @xmath75 . then , with using the distribution ( [ c ] ) , the condition ( [ 17 ] ) taken at @xmath76 arrives at the equipartition law @xmath77 according to which the value @xmath78 is the specific heat . let us stress that above scheme is related to generalized definition of the deformed logarithm @xcite @xmath79 where the function @xmath80 is reduced to the statistical weight distribution @xmath81 type of eq.([9 ] ) . however , this function is obeyed the condition @xmath82 which includes deformed temperature @xmath83 , being inconvenient at physical considerations . up to now , we suppose that the multifractal spectrum @xmath2 is arbitrary . if it is optimized at normalization condition @xmath84 one has to minimize the expression @xmath85^{-1}{\rm d}\gamma-\frac{\sigma ^2}{2}\left[\int\limits_0 ^ 1 f(d){\rm d}d-1\right ] , \label{20}\ ] ] written with accounting eqs.([2 ] ) , ( [ 8 ] ) , where @xmath86 is lagrange multiplier . as a result , we arrive at the equality @xmath87 - 2\tau(q)\right\}}\ln\gamma~{\rm d}\gamma=\frac{\sigma ^2}{2}\label{21}\ ] ] whose integration gives , with accounting eq.([9 ] ) , the transcendental equation @xmath88 ^ 2-w_q^{{\mathcal t } _ q(d)}\left[{\mathcal t}_q(d)\ln(w_q)-1\right]-1=0,\label{22}\\ { \mathcal t}_q(d)\equiv \bar{\tau}(q)\left\{1 - 2\tau(q)+\left[qd - f(d)\right]\right\}. \label{23}\end{aligned}\ ] ] this equation is written in the form , when can be used either given spectrum function @xmath2 or the index dependence @xmath34 . in the latter case , we find initially the dependence @xmath89 from the equation @xmath90 being conjugated to eq.([10 ] ) . then , substituting this dependence into eq.([23 ] ) , we arrive at the trivial expression @xmath91 whose using in eq.([22 ] ) allows us to determine the dependence of the statistical weight @xmath92 on the complexity @xmath93 at given function @xmath34 . with passage to the smooth phase space , when @xmath51 , @xmath25 , @xmath94 , one obtains the statistical weight @xmath95 which is determined by the specific complexity @xmath96 per one particle . at small deviation off the minimum complexity @xmath97 and light multifractality ( @xmath98 ) , linearized equation ( [ 22 ] ) gives @xmath99\left(\sigma -\sigma_\infty\right)\right\},\\ & w_{\tau(q)}\equiv\exp\left\{\tau(q)\sigma_\infty\left[1-\frac{1}{3}\left(1-\tau(q)\right ) \sigma_\infty\right]\right\},\quad 1\ll q<\infty . \end{split}\ ] ] in the opposite case @xmath100 , one has with logarithmic accuracy @xmath101^{\tau(q ) } , \ \alpha\sim 1,\quad q-1\ll 1 . \label{24a}\ ] ] as shows above consideration , the thermostatistics of complex systems with phase space , whose reduced fractal dimension @xmath0 is distributed with spectrum @xmath2 , is governed by the tsallis formalism of the non - extensive statistics . in this way , the role of non - additivity parameter plays inverted value of the multifractal function @xmath4 which monotonically increases , taking value @xmath37 at @xmath38 and @xmath102 at @xmath51 ( the latter limit is related to the smooth phase space ) . the multifractal function @xmath34 is reduced to the specific heat to determine , together with the inverted value @xmath3 , both statistical distributions and thermodynamic functions of the system under consideration . at given function @xmath34 , optimization of the normalized multifractal spectrum @xmath2 arrives at the dependence of the statistical weight on the system complexity .	we consider the self - similar phase space with reduced fractal dimension @xmath0 being distributed within domain @xmath1 with spectrum @xmath2 . related thermostatistics is shown to be governed by the tsallis formalism of the non - extensive statistics , where role of the non - additivity parameter plays inverted value @xmath3 of the multifractal function @xmath4 , being the specific heat , @xmath5 is multifractal parameter . in this way , the equipartition law is shown to take place . optimization of the multifractal spectrum @xmath2 derives the relation between the statistical weight and the system complexity . phase space ; fractal dimension spectrum ; deformed exponential 05.20.gg , 05.45.df , 05.70.ce .
here , we describe additional tests carried out in order to estimate the degree to which our dm limits might vary under alternative assumptions pertaining to the astrophysical background and cosmic ray propagation . in addition , we quantify the significance of spectral features in the observed positron fraction . in deriving our main results , as shown in fig . 3 , we used the phenomenological parameterization of the ams collaboration @xcite for the astrophysical contribution to the positron fraction , and adopted our reference assumptions of @xmath13 kpc and @xmath89 ev @xmath15 . in fig . [ fig : backpropvariation ] , for the case of direct dm annihilation to @xmath90 , we show in the _ left panel _ the impact of different propagation parameters when treating the astrophysical background in the same way as in fig . 3 . changing the diffusion conditions ( @xmath91 kpc ) in the galaxy in that case only affects our limits by @xmath92 , while allowing for higher energy losses ( @xmath93 ev @xmath15 ) can alter our limits by a factor of @xmath02 , with higher losses resulting in weaker limits ( see also fig . 1 ) . in the _ right panel _ , we repeat this exercise , but replace the ams background parametrization with physically motivated models for the primary @xmath94 , secondary @xmath2 , and pulsar originated @xmath2 fluxes ( see discussion in the main text ) , calculated with the same galactic propagation model as used in determining the spectrum of cr leptons from dm . in this case , our results can be further altered by a factor of up to @xmath03 . the reason for this change is that our physically motivated models describe the individual components by power - laws with breaks at a few gev . these spectral features in the background can be the result of different energy loss mechanisms kicking in , energy losses due to bremsstrahlung emission , dominant at lower energies , equal locally those due to synchrotron radiation and ics ( dominant at higher energies ) . since the energy loss rate @xmath95 due to bremsstrahlung radiation scales as @xmath96 while the @xmath95 due to synchrotron and ics as @xmath97 ( at the thompson cross - section regime ) , a spectral change in the propagated @xmath2 around that energy is expected ( see , e.g. , ref @xcite ) . ] or from individual local and recent supernovae affecting the high energy @xmath94 spectrum . also , observations at microwave and radio frequencies suggest a different spectral power - law for the cr @xmath2 at @xmath01gev @xcite compared to cr @xmath2 flux measurements at higher energies @xcite . while changes in the spectral power - law describing these components are motivated by the reasons just described , sharp breaks used to implement them are theoretically less accurate and fit slightly worse the ams positron fraction spectrum . in addition , our physically motivated models include the impact of solar modulation by using the force field approximation . solar modulation modifies the position and normalization of the dark matter signal flux , but is negligible at energies @xmath985gev . we do not expect solar modulation to significantly smoothen a sharp spectral peak at higher energies . given that we consider a population of pulsars as one possible source of the rising positron fraction above 10gev ( with tev - scale dm or a single dominant pulsar being alternative possibilities ) , we will briefly discuss the impact of their modeling on our limits . for pulsars that eventually inject equal amounts of @xmath2 into the ism , their injection spectra can be estimated from gamma - ray and synchrotron observations towards known pulsars , such as the crab . also get further accelerated at the termination shock between the magnetosphere and the pulsar wind nebulae ( pwn ) , and the @xmath2 injected to the ism are dominantly coming from middle aged pulsars after their respective pwn have been disrupted ( see refs . @xcite ) ] typical injection power - law values for the differential spectrum are expected to be in the range of 1 - 2 leading to propagated spectra with power - laws in the range of 2.0 @xmath6 0.5 . our fits for the averaged pulsar contribution agree with these expectations . final states , obtained by adopting the power - law background model . the _ gray lines _ , in contrast , show limits obtained when the contribution from many pulsars is taken into account ( for 15 different realizations ) . ] in addition , as suggested by refs . @xcite , the total contribution from many pulsars each with a different age , distance , initial rotational energy , injected energy into @xmath2 , and unique environmental surroundings affecting energy losses and diffusion is expected to give a spectrum with many peaks and dips , especially at higher energies where fewer pulsars significantly contribute . with fine enough energy resolution and high statistics , one should be able to observe such spectral features . by using the data from the atnf pulsar catalogue @xcite and implementing the parametrization of ref . @xcite , we ran multiple realizations of such combined spectra to study the impact of possible dips and peaks in the background spectrum on the derived dm limits . in particular , we include in these realizations all pulsars within 4kpc from us , except for millisecond pulsars and pulsars in binary systems . while we keep their individual locations and ages in all realizations fixed , we vary i ) the local cr diffusion properties and energy - losses , ii ) the cuts on the current spin - down power of pulsars as recorded in the atnf catalogue and , most importantly , iii ) the fraction @xmath99 of initial rotational energy of the individual neutron stars that is injected into the ism in the form of @xmath2 . we then fit to the ams data the injection spectral properties ( taken to be the same for all pulsars ) , the _ averaged _ value of @xmath99 , the primary sne @xmath94 , secondary @xmath2 cr flux normalizations and the solar modulation potential . even though the injection @xmath2 spectrum is taken to be the same for simplicity , however , the propagated pulsar spectra differ because of the different ages , distances and energy outputs ; this is clearly seen in the inset of fig . [ fig : limitsrand ] where we plot the resulting positron fraction . for dm channels that give broad continuous spectra , such as muons and taus ( see fig . 1 ) , the presence of multiple peaks and dips is unimportant . for hard spectra such as from monochromatic @xmath2 , however , our limits can be modified by a factor of up to @xmath03 , as also shown in fig . [ fig : limitsrand ] . in fig . [ fig : ts ] we show , for the case of @xmath48 final states , the local significance for a dm signal as function of the dm mass . the significance is plotted in units of gaussian sigma , and given by the square - root @xmath100 of the test statistics @xmath101 . here , @xmath102 denote respectively the likelihood of the alternative ( dm signal ) and null ( no dm signal ) hypothesis . for illustration , we also allow negative ( obviously unphysical ) signal normalizations in the fit , which are mapped onto the negative y - axis . as background model in the fit we use the reference power - law model from ref . we do not find any indications for local , edge - like , features in the ams data . lastly , as a simple cross - check , we have also run darksusy @xcite with standard parameters for propagation ( based on the prescription given in ref . @xcite ) and an nfw profile normalized to 0.4 gev @xmath15 . for the electron spectrum , we used a simple broken power law which agrees with the pamela electron data @xcite for @xmath103gev . knowing that the ams positron fraction measurement is well described by a simple background model , we then just demand that the dm signal does not exceed the reported @xmath104 error bars at the energy of the feature . the resulting limit curve agrees well with the more sophisticated treatment described in the main text .	the ams experiment onboard the international space station has recently provided cosmic ray electron and positron data with unprecedented precision in the range from 0.5 to 350gev . the observed rise in the positron fraction at energies above 10gev remains unexplained , with proposed solutions ranging from local pulsars to tev - scale dark matter . here , we make use of this high quality data to place stringent limits on dark matter with masses below @xmath0300gev , annihilating or decaying to leptonic final states , essentially independent of the origin of this rise . we significantly improve on existing constraints , in some cases by up to two orders of magnitude . [ [ introduction . ] ] introduction . + + + + + + + + + + + + + the ams ( alpha magnetic spectrometer ) collaboration has very recently announced the results of its first data collected from the international space station @xcite , consisting of a high precision measurement of the cosmic ray ( cr ) positron fraction @xcite . this new data provides a confirmation of the rise of this quantity above 10gev , as previously observed by pamela @xcite and fermi @xcite ( and with earlier hints provided by heat @xcite and ams-01 @xcite ) . such a rise is not predicted in the standard scenario , in which cr positrons are mostly produced as _ secondary _ particles , as a result of collisions of cr protons with the interstellar medium ( ism ) . instead , the large positron fraction seems to require the existence of at least one additional nearby _ primary _ source of high energy positrons . local pulsars have emerged as the leading astrophysical candidates @xcite , although it has also been argued that strong local sources might not actually be needed when taking into account the spiral structure of the milky way in full 3-d propagation models @xcite and that even a secondary production mechanism in the shock waves of supernovae remnants @xcite could provide a viable mechanism to explain the data @xcite . a more exotic possibility is that the observed positrons may be produced in the annihilations or decays of tev - scale dark matter ( dm ) particles . such scenarios , however , require unexpectedly large annihilation rates into predominantly leptonic final states @xcite and are subject to significant constraints from cr antiproton , gamma - ray and synchrotron data @xcite . upcoming ams data may help to settle this open issue not only by increasing statistics and extending their study to higher energies , but also by providing high precision measurements of other cr particle spectra ( likely breaking degeneracies in the propagation parameters @xcite ) . fermi and ams will also further constrain any anisotropy in the positron / electron flux ( where current limits are already close to discriminating between some of the scenarios described above @xcite ) . in this letter , we do not make any attempt to explain the origin of the rise in the positron fraction . instead , we focus on using the ams data to derive limits on subdominant exotic contributions to the observed cr positron spectrum , in particular from dm with masses below @xmath0300gev . while positrons have been used in the past to probe dm annihilation or decay @xcite , we exploit here for the first time the extremely high quality of the ams data to search for pronounced _ spectral features _ in the positron flux predicted in some dm models @xcite . much as exploiting spectral features can significantly improve the sensitivity of indirect dm searches using gamma rays @xcite , we demonstrate that the same is true for positrons , despite energy losses and other complicating factors . we derive limits that exceed the currently most stringent results on dm annihilation into leptons @xcite by up to two orders of magnitude . this letter is organized as follows . we first briefly review various astrophysical sources of leptons and how they manifest themselves in the observed cr flux , and then discuss possible contributions from dm . we continue with a description of the statistical treatment implemented here , before moving on to present our main results and conclusions . in an appendix @xcite , we collect further technical details of our procedure for deriving limits on a possible dm signal , discussing in particular the impact of systematic uncertainties in the background modeling . [ [ astrophysical - origins - of - cosmic - ray - leptons . ] ] astrophysical origins of cosmic ray leptons . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the origin of high energy electrons can be traced back to i ) supernova explosions that accelerate the ism to produce what are typically referred to as primary crs , ii ) inelastic collisions of primary cr protons and nuclei with the ism ( resulting in charged mesons , which decay , producing secondary electrons _ and _ positrons ) , and iii ) individual sources such as pulsars that produce @xmath1 pairs . the averaged spectrum of propagated primary cr electrons ( originating from many supernovae ) is expected to be harder than that of the secondary @xmath2 component because the primary cr progenitors of the secondaries have also experienced propagation effects ; both spectra are well described by power - laws , with spectral indices of about @xmath3 to @xmath4 ( @xmath5 ) for primary electrons ( secondary @xmath2 ) at energies above @xmath010gev @xcite . the contribution from all galactic pulsars can be approximated by a power - law with an exponential cut - off at high energies , with a propagated spectral index of 2.0 @xmath6 0.5 @xcite . the galactic magnetic field at scales @xmath7pc has a random and a regular component @xcite . as cr leptons propagate away from their sources , they follow the field lines and scatter off @xmath8-field irregularities . the net effect can be approximated as a random walk diffusion within a zone surrounding the galactic disk @xcite . further away from the disk the magnetic fields become weak , essentially leading to freely propagating crs . during their propagation throughout the galaxy , electrons and positrons also experience significant energy losses due to synchrotron and inverse compton scattering on the galactic radiation field and the cosmic microwave background . the impact of other effects such as convective winds , ionization losses , or positron annihilation in collisions with matter are not significant for leptons in the energy range considered here @xcite and are therefore ignored . we do , however , include bremsstrahlung emission , diffusive re - acceleration , and solar modulation inside the heliosphere ( using the force - field approximation @xcite ) , which have an impact on cr @xmath2 spectra below 510gev @xcite . for the propagation of cr leptons , we use the standard numerical tool galprop v54 @xcite , which includes up - to - date implementations of the local interstellar radiation field and galactic gas distribution . these are relevant for both the production of secondary leptons and energy losses . galprop assumes a diffusion zone with cylindrical symmetry within which crs diffuse and beyond which they escape . its scale height , @xmath9 , and other diffusion parameters , notably the diffusion time - scale and local diffusion properties , are constrained by observed cr ratios , including @xmath10 , b / c and @xmath11be/@xmath12be . as reference values we assume @xmath13 kpc , corresponding to the value best - fit by cr data @xcite and favored by radio observations @xcite , and the standard default galprop assumptions for the local radiation and magnetic field energy densities , corresponding to @xmath14ev@xmath15 @xcite . for the diffusion zone scale height , values of @xmath16kpc are in tension with a combined analysis of cr and gamma - ray data @xcite , while increasing @xmath9 beyond @xmath17kpc does not significantly alter our results . 2 , 4 , 8 kpc and rigidity @xmath18 , we adopt a diffusion coefficient @xmath19 , with @xmath20 0.81 , 1.90 , 2.65 ( @xmath21 @xmath22s@xmath23 ) , and alfvn velocities 9 , 10 , 10 km s@xmath23 . ] the propagation of high - energy leptons is actually dominated by energy losses rather than diffusion , implying that more conservative limits would arise for larger values of the local radiation and magnetic field energy densities . in our subsequent discussion , we will allow for an increase of @xmath24 by up to @xmath25 with respect to the reference value , which is still compatible with gamma - ray and synchrotron data @xcite . [ [ positrons - from - dark - matter . ] ] positrons from dark matter . + + + + + + + + + + + + + + + + + + + + + + + + + + + dm particles annihilating or decaying in the galactic halo may also contribute to the cr lepton spectrum , producing equal numbers of positrons and electrons . for annihilating dm , the injected spectrum of cr leptons per volume and time is given by @xmath26 ( divided by 2 if the dm particle is not self - conjugate ) , while for decaying dm , this is instead @xmath27 , where @xmath28 is the decay rate . here , @xmath29 is the velocity - averaged annihilation rate , @xmath30 is the dm density , @xmath31 is the dm mass , and @xmath32 is the spectrum of leptons produced per annihilation or decay . as our default choice , we adopt a dm distribution which follows an einasto profile @xcite , normalized to a local density of @xmath33gev @xcite . positrons from dm annihilation or decay typically result from the decay of @xmath34 ( for hadronic final states ) , or the leptonic decay of @xmath35 or @xmath36 . owing to the high multiplicity of such processes , the resulting @xmath37 energy distribution at injection ( which we take from ref . @xcite ) is typically very soft . if dm annihilates directly into @xmath1 , however , these are produced nearly monochromatically . even after accounting for energy losses from propagation , a very characteristic spectrum arises in this case , with a sharp edge - like feature at @xmath38 ( or at @xmath39 for decaying dm ) . a comparably distinct spectral feature arises from the annihilation of majorana dm into @xmath40 final states . popular examples for dm models with large annihilation rates into @xmath1 final states include kaluza - klein dm @xcite , while the supersymmetric neutralino is a possible candidate for producing a spectrum dominated by @xmath40 final states @xcite . we will always refer to that specific situation , dominated by photon emission off virtual selectrons @xmath41 . we assume at least one of the @xmath41 to be degenerate in mass with the neutralino . ] spectrum from annihilating dm , after propagation , for different annihilation final states , assuming @xmath42@xmath43 @xmath44s@xmath23 . solid lines refer to reference diffusion zone ( @xmath9=4kpc ) and energy loss assumptions ( @xmath45ev@xmath15 ) . dashed ( dotted ) lines show the effect of a different scale height @xmath9=8(2 ) kpc . the dash - dotted line shows the impact of increasing the local radiation plus magnetic field density to @xmath46ev@xmath15 . ] we illustrate this in fig . [ fig : spectra ] by showing the propagated @xmath1 spectra for various final states and an annihilation rate that corresponds to the thermal " cross section of @xmath47 ( which leads to the correct relic density in the simplest models of thermally produced dm ) . as anticipated , the @xmath48 and @xmath40 final states result in the most pronounced spectral features a fact which helps considerably , as we will see , to distinguish them from astrophysical backgrounds . for the case of @xmath48 final states , we also show how the spectrum depends on our local diffusion and energy loss assumptions within the range discussed above . increasing @xmath9 enables cr leptons to reach us from greater distances due to the larger diffusion volume and therefore results in softer propagated spectra . while the peak normalization of the spectrum depends only marginally on @xmath9 , it may be reduced by up to a factor of @xmath02 when increasing the assumed local energy losses via synchrotron radiation and inverse compton scattering by 50% . in fig . [ fig : fraction ] , we show a direct comparison of the dm signal with the ams data , for the case of @xmath48 final states contributing at the maximum level allowed by our constraints ( see below ) for two fiducial values of @xmath49 . again , it should be obvious that the shape of the dm contribution differs at all energies significantly from that of the background . , for @xmath50gev and 100gev . the normalization of the dm signal in each case was chosen such that it is barely excluded at the @xmath51 cl . for better visibility , the contribution from dm ( lower lines ) has been rescaled as indicated . ] [ [ statistical - treatment . ] ] statistical treatment . + + + + + + + + + + + + + + + + + + + + + + we use the likelihood ratio test @xcite to determine the significance of , and limits on , a possible dm contribution to the positron fraction measured by ams . as likelihood function , we adopt a product of normal distributions @xmath52 ; @xmath53 is the measured value , @xmath54 the positron fraction predicted by the model , and @xmath55 its variance . the dm contribution enters with a single degree of freedom , given by the non - negative signal normalization . upper limits at the @xmath51cl on the dm annihilation or decay rate are therefore derived by increasing the signal normalization from its best - fit value until @xmath56 is changed by 2.71 , while profiling over the parameters of the background model . we use data in the energy range 1350gev ; the variance @xmath55 is approximated by adding the statistical and systematic errors of the measurement in quadrature , @xmath57 . since the total relative error is always small ( below 17% ) , and at energies above 4gev dominated by statistics , we expect this approximation to be very reliable . the binning of the published positron fraction follows the ams energy resolution , which varies between 10.4% at 1gev and 1.5% at 350gev . although we do not account for the finite energy resolution of ams in our analysis , we have explicitly checked that this impacts our results by no more than 10% . as our nominal model for the part of the @xmath1 spectrum that does not originate from dm , henceforth simply referred to as the astrophysical background , we use the same phenomenological parameterization as the ams collaboration in their analysis @xcite . this parameterization describes each of the @xmath1 fluxes as the sum of a common source spectrum modeled as a power - law with exponential cutoff and an individual power - law contribution ( only the latter being different for the @xmath37 and @xmath58 fluxes ) . after adjusting normalization and slope of the secondary positrons such that the overall flux reproduces the fermi @xmath59 measurements @xcite , the five remaining model parameters are left unconstrained . this phenomenological parameterization provides an extremely good fit ( with a @xmath60 ) , indicating that no fine structures are observed in the ams data . for the best - fit spectral slopes of the individual power - laws we find @xmath61 and @xmath62 , respectively , and for the common source @xmath63 with a cutoff at @xmath64800gev , consistent with ref . @xcite . subsequently , we will keep @xmath65 fixed to its best - fit value . [ [ results - and - discussion . ] ] results and discussion . + + + + + + + + + + + + + + + + + + + + + + + cl ) on the dm annihilation cross section , as derived from the ams positron fraction , for various final states ( this work ) , wmap7 ( for @xmath66 ) @xcite and fermi lat dwarf spheroidals ( for @xmath67 and @xmath68 ) @xcite . the dotted portions of the curves are potentially affected by solar modulation . we also indicate @xmath69 . the ams limits are shown for reasonable reference values of the local dm density and energy loss rate ( see text ) , and can vary by a factor of a few , as indicated by the hatched band ( for clarity , this band is only shown around the @xmath70 constraint ) . ] our main results are the bounds on the dm annihilation cross section , as shown in fig . [ fig : results ] . no significant excess above background was observed . for annihilations proceeding entirely to @xmath48 final states , we find that the `` thermal '' cross section is firmly excluded for @xmath71gev . for @xmath72gev , which is an interesting range in light of recent results from direct @xcite and indirect @xcite dm searches , our upper bound on the annihilation cross section to @xmath70 is approximately two orders of magnitude below @xmath73 . if only a fraction @xmath74 of dm annihilates like assumed , limits would scale like @xmath75 ( and , very roughly , @xmath76 ) . we also show in fig . [ fig : results ] the upper bounds obtained for other leptonic final states . as expected , these limits are weaker than those found in the case of direct annihilation to electrons both because part of the energy is taken away by other particles ( neutrinos , in particular ) and because they feature broader and less distinctive spectral shapes . these new limits on dm annihilating to @xmath67 and @xmath68 final states are still , however , highly competitive with or much stronger than those derived from other observations , such as from the cosmic microwave background @xcite and from gamma - ray observations of dwarf galaxies @xcite . note that for the case of @xmath40 final states even stronger limits can be derived for @xmath77gev by a spectral analysis of gamma rays @xcite . we do not show results for the @xmath78 channel , for which we nominally find even weaker limits due to the broader spectrum ( for @xmath79gev , about @xmath80 ) . in fact , due to degeneracies with the background modeling , limits for annihilation channels which produce such a broad spectrum of positrons can suffer from significant systematic uncertainties . for this reason , we consider our limits on the @xmath48 channel to be the most robust . uncertainties in the @xmath2 energy loss rate and local dm density weaken , to some extent , our ability to robustly constrain the annihilation cross sections under consideration in fig . [ fig : results ] . we reflect this uncertainty by showing a band around the @xmath48 constraint , corresponding to the range @xmath81ev@xmath15 , and @xmath82gev@xmath15 @xcite ( note that the _ form _ of the dm profile has a much smaller impact ) . uncertainty bands of the same width apply to each of the other final states shown in the figure , but are not explicitly shown for clarity . other diffusion parameter choices impact our limits only by up to @xmath010% , except for the case of low dm masses , for which the effect of solar modulation may be increasingly important @xcite . we reflect this in fig . [ fig : results ] by depicting the limits derived in this less certain mass range , where the peak of the signal @xmath37 flux ( as shown in fig . [ fig : spectra ] ) falls below a fiducial value of 5gev , with dotted rather than solid lines . for comparison , we have also considered a collection of physical background models in which we calculated the expected primary and secondary lepton fluxes using galprop , and then added the contribution from all galactic pulsars . while this leads to an almost identical description of the background at high energies as in the phenomenological model , small differences are manifest at lower energies due to solar modulation and a spectral break @xcite in the cr injection spectrum at a few gev ( both neglected in the ams parameterization ) . we cross - check our fit to the ams positron fraction with lepton measurements by fermi @xcite . using these physical background models in our fits , instead of the phenomenological ams parameterization , the limits do not change significantly . the arguably most extreme case would be the appearance of dips in the background due to the superposition of several pulsar contributions , which might conspire with a hidden dm signal at almost exactly the same energy . we find that in such situations , the real limits on the annihilation rate could be weaker ( or stronger ) by up to roughly a factor of 3 for any individual value of @xmath49 . see the appendix @xcite for more details and further discussion of possible systematics that might affect our analysis . lastly , we note that the upper limits on @xmath83 reported in fig . [ fig : results ] can easily be translated into upper limits on the decay width of a dm particle of mass @xmath84 via @xmath85 . we checked explicitly that this simple transformation is correct to better than 10% for the @xmath864 kpc propagation scenario and @xmath48 and @xmath87 final states over the full considered energy range . [ [ conclusions . ] ] conclusions . + + + + + + + + + + + + in this letter , we have considered a possible dark matter contribution to the recent ams cosmic ray positron fraction data . the high quality of this data has allowed us for the first time to successfully perform a spectral analysis , similar to that used previously in the context of gamma ray searches for dm . while we have found no indication of a dm signal , we have derived upper bounds on annihilation and decay rates into leptonic final states that improve upon the most stringent current limits by up to two orders of magnitude . for light dm in particular , our limits for @xmath48 and @xmath87 final states are significantly below the cross section naively predicted for a simple thermal relic . when taken together with constraints on dm annihilations to hadronic final states from gamma rays @xcite and antiprotons @xcite , this new information significantly limits the range of models which may contain a viable candidate for dark matter with @xmath88gev . the ams mission is planned to continue for 20 years . compared to the 18 months of data @xcite our analysis is based on , we expect to be able to strengthen the presented limits by at least a factor of three in the energy range of 6200gev with the total data set , and by more in the likely case that systematics and the effective acceptance of the instrument improve . this work makes use of scipy @xcite , minuit @xcite and matplotlib @xcite . the research of l.b . was carried out under swedish research council ( vr ) contract no . 621 - 2009 - 3915 . t.b . acknowledges support from the german research foundation ( dfg ) through the emmy noether grant br 3954/1 - 1 . i.c . , c.w . and d.h . thank the _ kavli institute for theoretical physics _ in santa barbara , california , for their kind hospitality . this work has been supported by the us department of energy . 00 m. aguilar _ et al . _ [ ams collaboration ] , phys . rev . lett . * 110 * , no . 14 , 141102 ( 2013 ) . o. adriani _ et al . _ [ pamela collaboration ] , nature * 458 * , 607 ( 2009 ) [ arxiv:0810.4995 [ astro - ph ] ] . m. ackermann _ et al . _ [ fermi lat collaboration ] , phys . rev . lett . * 108 * , 011103 ( 2012 ) [ arxiv:1109.0521 [ astro-ph.he ] ] . s. w. barwick _ et al . _ [ heat collaboration ] , astrophys . j. * 482 * , l191 ( 1997 ) [ astro - ph/9703192 ] . m. aguilar _ et al . _ [ ams-01 collaboration ] , phys . lett . b * 646 * , 145 ( 2007 ) [ astro - ph/0703154 [ astro - ph ] ] . d. hooper , p. blasi and p. d. serpico , jcap * 0901 * , 025 ( 2009 ) [ arxiv:0810.1527 [ astro - ph ] ] . s. profumo , central eur . j. phys . * 10 * , 1 ( 2011 ) [ arxiv:0812.4457 [ astro - ph ] ] . d. gaggero , l. maccione , g. di bernardo , c. evoli and d. grasso , phys . rev . lett . * 111 * , 021102 ( 2013 ) [ arxiv:1304.6718 [ astro-ph.he ] ] . p. blasi , phys . rev . lett . * 103 * , 051104 ( 2009 ) [ arxiv:0903.2794 [ astro-ph.he ] ] . p. mertsch and s. sarkar , phys . rev . lett . * 103 * , 081104 ( 2009 ) [ arxiv:0905.3152 [ astro-ph.he ] ] . for a review , see p. d. serpico , astropart . phys . * 39 - 40 * , 2 ( 2012 ) [ arxiv:1108.4827 [ astro-ph.he ] ] . l. bergstrm , j. edsj and g. zaharijas , phys . rev . lett . * 103 * , 031103 ( 2009 ) [ arxiv:0905.0333 [ astro-ph.he ] ] . d. p. finkbeiner , l. goodenough , t. r. slatyer , m. vogelsberger and n. weiner , jcap * 1105 * , 002 ( 2011 ) [ arxiv:1011.3082 [ hep - ph ] ] . q. yuan , x. -j . bi , g. -m . chen , y. -q . guo , s. -j . lin and x. zhang , arxiv:1304.1482 [ astro-ph.he ] . i. cholis and d. hooper , phys . rev . d * 88 * , 023013 ( 2013 ) [ arxiv:1304.1840 [ astro-ph.he ] ] . h. -b . jin , y. -l . wu and y. -f . zhou , arxiv:1304.1997 [ hep - ph ] . f. donato , d. maurin , p. brun , t. delahaye and p. salati , phys . rev . lett . * 102 * , 071301 ( 2009 ) [ arxiv:0810.5292 [ astro - ph ] ] . g. bertone , m. cirelli , a. strumia and m. taoso , jcap * 0903 * , 009 ( 2009 ) [ arxiv:0811.3744 [ astro - ph ] ] . l. bergstrm , g. bertone , t. bringmann , j. edsj and m. taoso , phys . rev . d * 79 * , 081303 ( 2009 ) [ arxiv:0812.3895 [ astro - ph ] ] . m. cirelli , p. panci and p. d. serpico , nucl . phys . b * 840 * , 284 ( 2010 ) [ arxiv:0912.0663 [ astro-ph.co ] ] . c. evoli , i. cholis , d. grasso , l. maccione and p. ullio , phys . rev . d * 85 * , 123511 ( 2012 ) [ arxiv:1108.0664 [ astro-ph.he ] ] . r. kappl and m. w. winkler , phys . rev . d * 85 * , 123522 ( 2012 ) [ arxiv:1110.4376 [ hep - ph ] ] . m. wechakama and y. ascasibar , arxiv:1212.2583 [ astro-ph.co ] . a. de simone , a. riotto and w. xue , jcap * 1305 * , 003 ( 2013 ) [ arxiv:1304.1336 [ hep - ph ] ] . m. pato , d. hooper and m. simet , jcap * 1006 * , 022 ( 2010 ) [ arxiv:1002.3341 [ astro-ph.he ] ] . m. ackermann _ et al . _ [ fermi - lat collaboration ] , phys . rev . d * 82 * , 092003 ( 2010 ) [ arxiv:1008.5119 [ astro-ph.he ] ] . g. di bernardo , c. evoli , d. gaggero , d. grasso , l. maccione and m. n. mazziotta , astropart . phys . * 34 * , 528 ( 2011 ) [ arxiv:1010.0174 [ astro-ph.he ] ] . j. r. ellis , r. a. flores , k. freese , s. ritz , d. seckel and j. silk , phys . lett . b * 214 * , 403 ( 1988 ) . s. rudaz and f. w. stecker , astrophys . j. * 325 * , 16 ( 1988 ) . m. kamionkowski and m. s. turner , phys . rev . d * 43 * , 1774 ( 1991 ) . e. a. baltz and j. edsj , phys . rev . d * 59 * , 023511 ( 1998 ) [ astro - ph/9808243 ] . e. a. baltz , j. edsj , k. freese and p. gondolo , phys . rev . d * 65 * , 063511 ( 2002 ) [ astro - ph/0109318 ] . j. kopp , arxiv:1304.1184 [ hep - ph ] . a. j. tylka , phys . rev . lett . * 63 * , 840 ( 1989 ) [ erratum - ibid . * 63 * , 1658 ( 1989 ) ] . m. s. turner and f. wilczek , phys . rev . d * 42 * , 1001 ( 1990 ) . e. a. baltz and l. bergstrm , phys . rev . d * 67 * , 043516 ( 2003 ) [ hep - ph/0211325 ] . d. hooper and g. d. kribs , phys . rev . d * 70 * , 115004 ( 2004 ) [ hep - ph/0406026 ] . e. a. baltz and d. hooper , jcap * 0507 * , 001 ( 2005 ) [ hep - ph/0411053 ] . l. bergstrm , t. bringmann and j. edsj , phys . rev . d * 78 * , 103520 ( 2008 ) [ arxiv:0808.3725 [ astro - ph ] ] . d. hooper and w. xue , phys . rev . lett . * 110 * , 041302 ( 2013 ) [ arxiv:1210.1220 [ astro-ph.he ] ] . t. bringmann , f. calore , g. vertongen and c. weniger , phys . rev . d * 84 * , 103525 ( 2011 ) [ arxiv:1106.1874 [ hep - ph ] ] . m. ackermann _ et al . _ [ fermi - lat collaboration ] , phys . rev . lett . * 107 * , 241302 ( 2011 ) [ arxiv:1108.3546 [ astro-ph.he ] ] . s. galli , f. iocco , g. bertone and a. melchiorri , phys . rev . d * 80 * , 023505 ( 2009 ) [ arxiv:0905.0003 [ astro-ph.co ] ] ; t. r. slatyer , n. padmanabhan and d. p. finkbeiner , phys . rev . d * 80 * , 043526 ( 2009 ) [ arxiv:0906.1197 [ astro-ph.co ] ] ; s. galli , f. iocco , g. bertone and a. melchiorri , phys . rev . d * 84 * , 027302 ( 2011 ) [ arxiv:1106.1528 [ astro-ph.co ] ] ; t. r. slatyer , arxiv:1211.0283 [ astro-ph.co ] ; j. m. cline and p. scott , jcap * 1303 * , 044 ( 2013 ) [ erratum - ibid . * 1305 * , e01 ( 2013 ) ] [ arxiv:1301.5908 [ astro-ph.co ] ] ; l. lopez - honorez , o. mena , s. palomares - ruiz and a. c. vincent , jcap * 1307 * , 046 ( 2013 ) [ arxiv:1303.5094 [ astro-ph.co ] ] . see appendix . see , e.g. , i. v. moskalenko and a. w. strong , astrophys . j. * 493 * , 694 ( 1998 ) [ astro - ph/9710124 ] ; r. schlickeiser , `` cosmic ray astrophysics '' springer ( 2002 ) , and references therein . for an alternative discussion , see also b. katz , k. blum and e. waxman , mon . not . roy . astron . soc . * 405 * , 1458 ( 2010 ) [ arxiv:0907.1686 [ astro - ph ] ] . see r. jansson , g. r. farrar , a. h. waelkens and t. a. ensslin , jcap * 0907 * , 021 ( 2009 ) [ arxiv:0905.2228[astro - ph ] ] . and references therein . m. s. longair , `` high energy astrophysics '' , vol . * 1 * pp . 346 - 350 , cambridge university press ( 2004 ) . a. w. strong , i. v. moskalenko and v. s. ptuskin , ann . rev . nucl . part . sci . * 57 * , 285 ( 2007 ) [ astro - ph/0701517 ] . m. s. longair , `` high energy astrophysics '' , vol . * 2 * pp . 274 - 286 , cambridge university press ( 2004 ) . l. j. gleeson and w. i. axford , astrophys . j. * 154 * , 1011 ( 1968 ) . r. d. strauss , m. s. potgieter , i. bsching and a. kopp , astrophys . j. * 735 * , 83 ( 2011 ) . l. maccione , phys . rev . lett . * 110 * , 081101 ( 2013 ) [ arxiv:1211.6905[astro - ph.he ] ] . a. w. strong , i. v. moskalenko , t. a. porter , e. orlando , s. w. digel and a. e. vladimirov `` galprop v.54 : explanatory supplement ( 2011 ) '' ; http://galprop.stanford.edu/ . r. trotta , g. johannesson , i. v. moskalenko , t. a. porter , r. r. de austri and a. w. strong , astrophys . j. * 729 * , 106 ( 2011 ) [ arxiv:1011.0037[astro - ph.he ] ] . t. bringmann , f. donato and r. a. lineros , jcap * 1201 * , 049 ( 2012 ) [ arxiv:1106.4821 [ astro-ph.ga ] ] . i. cholis , m. tavakoli , c. evoli and p. ullio , jcap * 1205 * , 004 ( 2012 ) [ arxiv:1106.5073[astro - ph.he ] ] . t. r. jaffe , a. j. banday , j. p. leahy , s. leach and a. w. strong , [ arxiv:1105.5885[astro - ph.ga ] ] . d. merritt , j. f. navarro , a. ludlow and a. jenkins , astrophys . j. * 624 * , l85 ( 2005 ) [ astro - ph/0502515 ] . r. catena and p. ullio , jcap * 1008 * , 004 ( 2010 ) [ arxiv:0907.0018 [ astro-ph.co ] ] . p. salucci , f. nesti , g. gentile and c. f. martins , astron . astrophys . * 523 * , a83 ( 2010 ) [ arxiv:1003.3101 [ astro-ph.ga ] ] . m. cirelli , g. corcella , a. hektor , g. hutsi , m. kadastik , p. panci , m. raidal and f. sala _ et al . _ , jcap * 1103 * , 051 ( 2011 ) [ erratum - ibid . * 1210 * , e01 ( 2012 ) ] [ arxiv:1012.4515 [ hep - ph ] ] . w. a. rolke , a. m. lopez and j. conrad , nucl . instrum . meth . a * 551 * , 493 ( 2005 ) [ physics/0403059 ] . m. ackermann _ et al . _ [ fermi lat collaboration ] , phys . rev . d * 82 * , 092004 ( 2010 ) [ arxiv:1008.3999 [ astro-ph.he ] ] . r. bernabei _ et al . _ [ dama and libra collaborations ] , eur . phys . j. c * 67 * , 39 ( 2010 ) [ arxiv:1002.1028 [ astro-ph.ga ] ] . c. e. aalseth , p. s. barbeau , j. colaresi , j. i. collar , j. diaz leon , j. e. fast , n. fields and t. w. hossbach _ et al . _ , phys . rev . lett . * 107 * , 141301 ( 2011 ) [ arxiv:1106.0650 [ astro-ph.co ] ] . g. angloher , m. bauer , i. bavykina , a. bento , c. bucci , c. ciemniak , g. deuter and f. von feilitzsch _ et al . _ , eur . phys . j. c * 72 * , 1971 ( 2012 ) [ arxiv:1109.0702 [ astro-ph.co ] ] . r. agnese _ et al . _ [ cdms collaboration ] , [ arxiv:1304.4279 [ hep - ex ] ] . m. t. frandsen , f. kahlhoefer , c. mccabe , s. sarkar and k. schmidt - hoberg , jcap * 1307 * , 023 ( 2013 ) [ arxiv:1304.6066 [ hep - ph ] ] . d. hooper and t. linden , phys . rev . d * 84 * , 123005 ( 2011 ) [ arxiv:1110.0006 [ astro-ph.he ] ] . d. hooper and t. r. slatyer , arxiv:1302.6589 [ astro-ph.he ] . k. n. abazajian and m. kaplinghat , phys . rev . d * 86 * , 083511 ( 2012 ) [ arxiv:1207.6047 [ astro-ph.he ] ] . t. bringmann , x. huang , a. ibarra , s. vogl and c. weniger , jcap * 1207 * , 054 ( 2012 ) [ arxiv:1203.1312 [ hep - ph ] ] . f. iocco , m. pato , g. bertone and p. jetzer , jcap * 1111 * , 029 ( 2011 ) [ arxiv:1107.5810 [ astro-ph.ga ] ] . s. della torre , p. bobik , m. j. boschini , c. consolandi , m. gervasi , d. grandi , k. kudela and s. pensotti _ et al . _ , adv . space res . * 49 * , 1587 ( 2012 ) . a. w. strong , i. v. moskalenko and o. reimer , astrophys . j. * 537 * , 763 ( 2000 ) [ erratum - ibid . * 541 * , 1109 ( 2000 ) ] [ astro - ph/9811296 ] ; a. w. strong , i. v. moskalenkoand o. reimer , astrophys . j. * 613 * , 962 ( 2004 ) [ astro - ph/0406254 ] . v. s. ptuskin , i. v. moskalenko , f. c. jones , a. w. strong and v. n. zirakashvilli , astrophys . j. * 642 * , 902 ( 2006 ) [ astro - ph/0510335 ] . e. jones , t. oliphant , p. peterson , _ et . al . _ , _ scipy : open source scientific tools for python _ , 2001 . f. james and m. roos , comput . phys . commun . * 10 * , 343 ( 1975 ) . j. d. hunter , matplotlib : a 2d graphics environment , comput . in science & engineering * 9 * , 90 ( 2007 ) . o. adriani , _ et al . _ [ pamela collaboration ] , phys . rev . lett . * 106 * , 201101 ( 2011 ) . r. d. strauss , m. s. potgieter , i. bsching and a. kopp , astrophys . & sp . sc . * 339 * , 223 ( 2012 ) . j. arons , space sci . rev . * 75 * , 235 ( 1996 ) [ arxiv:0811.3894 [ astro - ph ] ] . d. a. green , [ arxiv:0905.3699 [ astro - ph ] ] . d. malyshev , i. cholis and j. gelfand , phys . rev . d * 80 * , 063005 ( 2009 ) [ arxiv:0903.1310 [ astro-ph.he ] ] . t. kobayashi , y. komori , k. yoshida and j. nishimura , astrophys . j. * 601 * , 340 ( 2004 ) [ astro - ph/0308470 [ astro - ph ] ] . d. grasso _ et al . _ [ fermi - lat collaboration ] , astropart . phys . * 32 * , 140 ( 2009 ) [ arxiv:0905.0636 [ astro-ph.he ] ] . r. n. manchester , g. b. hobbs , a. teoh and m. hobbs , astrophys . j. * 129 * , 1993 ( 2005 ) ; http://www.atnf.csiro.au/people/pulsar/psrcat/ . p. gondolo , j. edsj , p. ullio , l. bergstrm , m. schelke and e. a. baltz , jcap * 0407 * , 008 ( 2004 ) [ astro - ph/0406204 ] ; p. gondolo , j. edsj , l. bergstrm , p. ullio , m. schelke , e. a. baltz , t. bringmann and g. duda , http://www.darksusy.org .
in this paper , we have investigated the dynamics of reheating and thermalization of light particles which are produced from the perturbative decay of the scalar field , especially paying attention to the case with a small decay width of the scalar field , such as the one given by planck - suppressed interactions . since the number density is under occupied in this case , number violating splittings with small momentum exchanges and the evolution of the soft sector play crucial roles in the dynamics towards thermal equilibrium . importantly , the bottleneck process is the dissipation of remaining hard particles into the soft sector as pointed out in refs . @xcite in the context of qgp . following their arguments , we have studied thermalization after / during reheating in detail , while paying attention to the red - shift of previously produced particles before the completion of reheating and to the dilution of their number density by the cosmic expansion . we have shown that thermalization takes place instantaneously in most cases , and also found that red - shifted previously produced particles are not relevant since radiated soft particles dominate the number density of red - shifted previously produced ones and that radiation is always dominated by that from hardest primaries which are not red - shifted . the condition under which produced particles thermalize before the completion of reheating is given by @xmath3 which turns out to be much milder than that obtained in the previous work @xcite because the growth of the soft sector promotes the dissipation of remaining hard particles . as a result , even if the mass of the scalar field is as large as @xmath4 , it is shown that thermalization takes place instantaneously in most cases . our result is applicable to general situations where a decay of a non - relativistic matter reheats the universe via a small decay width and the decay products contain particles that are charged under a non - abelian gauge group . our analysis can also be applied to the instant preheating scenario @xcite in which produced light particles have small number densities compared with thermal ones . the situation is quite similar to the case we have considered and hence our result is helpful in estimating the time scale when non - perturbative particle production is shut off by rescatterings @xcite ( see also @xcite ) . the authors thank shigeki matsumoto and kazunori nakayama for useful discussions . this work is supported by the world premier international research center initiative ( wpi initiative ) , mext , japan ( k.h . ) , and jsps research fellowships for young scientists ( k.h . and k.m . ) . a. a. penzias and r. w. wilson , astrophys . j. * 142 * , 419 ( 1965 ) . j. c. mather , e. s. cheng , d. a. cottingham , r. e. eplee , d. j. fixsen , t. hewagama , r. b. isaacman and k. a. jesnsen _ et al . _ , astrophys . j. * 420 * , 439 ( 1994 ) . d. n. schramm and m. s. turner , rev . phys . * 70 * , 303 ( 1998 ) [ astro - ph/9706069 ] , and references therein . a. h. guth , phys . d * 23 * , 347 ( 1981 ) ; see also a. a. starobinsky , phys . b * 91 * , 99 ( 1980 ) ; k. sato , mon . not . soc . * 195 * , 467 ( 1981 ) . a. d. linde and v. f. mukhanov , phys . d * 56 * , 535 ( 1997 ) [ astro - ph/9610219 ] ; k. enqvist and m. s. sloth , nucl . b * 626 * , 395 ( 2002 ) [ hep - ph/0109214 ] ; d. h. lyth and d. wands , phys . b * 524 * , 5 ( 2002 ) [ hep - ph/0110002 ] ; t. moroi and t. takahashi , phys . b * 522 * , 215 ( 2001 ) [ erratum - ibid . b * 539 * , 303 ( 2002 ) ] [ hep - ph/0110096 ] . g. d. coughlan , w. fischler , e. w. kolb , s. raby and g. g. ross , phys . b * 131 * , 59 ( 1983 ) ; t. banks , d. b. kaplan and a. e. nelson , phys . rev . d * 49 * , 779 ( 1994 ) [ hep - ph/9308292 ] ; b. de carlos , j. a. casas , f. quevedo and e. roulet , phys . b * 318 * , 447 ( 1993 ) [ hep - ph/9308325 ] . m. fukugita and t. yanagida , phys . b * 174 * , 45 ( 1986 ) . w. buchmuller , p. di bari and m. plumacher , annals phys . * 315 * , 305 ( 2005 ) [ hep - ph/0401240 ] . m. kawasaki and t. moroi , prog . phys . * 93 * , 879 ( 1995 ) [ hep - ph/9403364 , hep - ph/9403061 ] ; m. kawasaki , k. kohri , t. moroi and a. yotsuyanagi , phys . d * 78 * , 065011 ( 2008 ) [ arxiv:0804.3745 [ hep - ph ] ] . m. dine , l. randall and s. d. thomas , nucl . b * 458 * , 291 ( 1996 ) [ hep - ph/9507453 ] ; r. allahverdi , b. a. campbell and j. r. ellis , nucl . b * 579 * , 355 ( 2000 ) [ hep - ph/0001122 ] . a. anisimov and m. dine , nucl . b * 619 * , 729 ( 2001 ) [ hep - ph/0008058 ] ; m. fujii , k. hamaguchi and t. yanagida , phys . d * 63 * , 123513 ( 2001 ) [ hep - ph/0102187 ] . j. r. ellis , k. enqvist , d. v. nanopoulos and k. a. olive , phys . b * 191 * , 343 ( 1987 ) . j. mcdonald , phys . d * 61 * , 083513 ( 2000 ) [ hep - ph/9909467 ] . r. allahverdi , phys . d * 62 * , 063509 ( 2000 ) [ hep - ph/0004035 ] . a. mazumdar and b. zaldivar , arxiv:1310.5143 [ hep - ph ] . s. dimopoulos , g. r. dvali and r. rattazzi , phys . b * 410 * , 119 ( 1997 ) [ hep - ph/9705348 ] . f. takahashi , phys . b * 693 * , 140 ( 2010 ) [ arxiv:1006.2801 [ hep - ph ] ] ; k. nakayama and f. takahashi , jcap * 1011 * , 009 ( 2010 ) [ arxiv:1008.2956 [ hep - ph ] ] . k. harigaya , m. ibe , k. schmitz and t. t. yanagida , phys . b * 720 * , 125 ( 2013 ) [ arxiv:1211.6241 [ hep - ph ] ] . j. i . yokoyama , phys . d * 70 * , 103511 ( 2004 ) [ hep - ph/0406072 ] ; j. i . yokoyama , phys . b * 635 * , 66 ( 2006 ) [ hep - ph/0510091 ] . m. bastero - gil , a. berera and r. o. ramos , jcap * 1109 * , 033 ( 2011 ) [ arxiv:1008.1929 [ hep - ph ] ] ; m. bastero - gil , a. berera , r. o. ramos and j. g. rosa , jcap * 1301 * , 016 ( 2013 ) [ arxiv:1207.0445 [ hep - ph ] ] . m. drewes , arxiv:1012.5380 [ hep - th ] . l. d. landau and i. pomeranchuk , dokl . nauk ser . * 92 * , 535 ( 1953 ) ; a. b. migdal , phys . rev . * 103 * , 1811 ( 1956 ) . b. arnold , g. d. moore and l. g. yaffe , jhep * 0111 * , 057 ( 2001 ) [ hep - ph/0109064 ] ; p. b. arnold , g. d. moore and l. g. yaffe , jhep * 0112 * , 009 ( 2001 ) [ hep - ph/0111107 ] . b. arnold , g. d. moore and l. g. yaffe , jhep * 0206 * , 030 ( 2002 ) [ hep - ph/0204343 ] . d. besak and d. bodeker , jhep * 1005 * , 007 ( 2010 ) [ arxiv:1002.0022 [ hep - ph ] ] .	if reheating of the universe takes place via planck - suppressed decay , it seems that the thermalization of produced particles might be delayed , since they have large energy / small number densities and number violating large angle scatterings which decrease the momentum of particles by large amount are inefficient correspondingly . in this paper , we study the thermalization of such `` under occupied '' decay products in detail , following recent developments in understanding the thermalization of non - abelian plasma . contrary to the above naive expectation , it is shown that in most cases thermalization after / during reheating occurs instantaneously by properly taking account of scatterings with small angles and of particles with small momenta . in particular , the condition for instantaneous thermalization before the completion of reheating is found to be @xmath0 , which is much milder than that obtained in previous works with small angle scatterings taken into account . c u v term_\|^2 addtoresetequationsection ipmu 13 - 0233 + ut-13 - 43 * thermalization after / during reheating + * .65 in keisuke harigaya@xmath1 and kyohei mukaida@xmath2 .4 in [ cols= " < , < " , ] [ tab : time ]
the experiment was performed on a @xmath20 heterostructure grown by molecular - beam epitaxy , with a 90-nm - deep 2deg with an electron density of @xmath21 and mobility of @xmath22 ( measured at 1.3 k ) . the metallic ( ti - au ) surface gates were fabricated using electron - beam lithography . the device was cooled inside an oxford triton 400 dilution refrigerator to a base temperature of 15 mk . to reduce charge noise the sample was cooled while applying a positive voltage on all gates ( ranging between 100 and 400 mv ) @xcite . gates @xmath5 , @xmath6 and @xmath7 were connected to homebuilt bias - tees ( @xmath23=470 ms ) , enabling application of d.c . voltage bias as well as high - frequency voltage excitation to these gates . frequency multiplexing combined with rf reflectometry of the sds was performed using two lc circuits matching a carrier wave of frequency 107.1 mhz for sd1 and 86.4 mhz for sd2 . the inductors are formed by microfabricated nbtin superconducting spiral inductors with an inductance of 3.2 @xmath24h ( sd1 ) and 4.6 @xmath24h ( sd2 ) . the power of the carrier wave arriving at the sample was estimated to be -93 dbm . the reflected signal was amplified using a cryogenic weinreb citlf2 amplifier and subsequently demodulated using homebuilt electronics . data acquisition was performed using a fpga ( de0-nano terasic ) and digital multimeters ( keithley ) . voltage pulses to the gates were applied using a tektronix awg5014 . the software was developed using python @xcite with scipy @xcite . the image processing is performed in pixel coordinates . we specify the parameters of algorithms in physical units such as mv . the corresponding parameter in pixel units is then determined by translating the value using the scan parameters . by specifying the parameters in physical units the algorithms remain valid if we make scans with a different resolution . of course making scans with a different resolution can lead to differences in rounding of numbers leading to slightly different results . * determine the lowvalue ( @xmath1 ) and highvalue ( @xmath25 ) of the scan by taking a robust minimum and maximum . for @xmath1 this is done by taking the 1^th^ percentile of the values . @xmath25 is determined by first taking the 90^th^ percentile of the scan data @xmath26 and then the 90^th^ percentile of all the values larger then ( @xmath27)/2 . this two - stage process to determine @xmath25 also works well when the pinch - off occurs for very positive gate voltages . simply taking for example the 99^th^ percentile of the scan data could then result in a too low estimate . * smoothen the signal and find the first element in the scan larger than @xmath28 . the position of that value is selected as the transition value . * perform several additional checks . the above two steps will always results in a transition value , even though the channel could be completely open or closed . the checks include amongst others : * * if the transition value is near the left border of the scan we conclude the transition has not been reached . we then set the transition value to the lowest value of the gate that has been scanned . in principle the algorithm could continue to search for a transition value for more negative gate voltages . however , making gate voltages too negative may induce charge noise in the sample so we do not want to apply very negative voltages . choosing the most negative voltage of the scan range then turns out to be a good choice . in the next steps of the algorithm , this transition voltage is just a starting value and the gate voltage will still be varied . due to cross - talk , the neighboring gates in follow - up steps will together with the gate that did not yet close , typically still ensure the formation of single dots . * * the difference of the mean of the measured values left of the transition value and the mean of the values right of the transition value should be large enough . large enough means more than 0.3 times the standard deviation of all the values in the scan . if it is not large enough , we set the transition value to the lowest value of the gate that has been scanned following a similar reasoning as for the previous check . in this scenario we assume that the scan range started at a voltage around 0 mv , and thus that no significant change in the measured current corresponds to a channel that was always open . as described in the main text the initial analysis of a 2d scan is performed by fitting a tetragon to the image . the bottom - left corner point of this tetragon gives a good indication of the position of the area where coulomb peaks are visible . see the magenta points in figs . [ fig : quadrilateral_fit](a - d ) . the important structures in the scan of a double dot are the charging lines and the crossings of two charging lines . to determine the locations of the crossings in the image we create a template for such a crossing . we then search for crossings using template matching . the response to the template is thresholded and local maxima are calculated . the template of the crossing consists of 4 lines at angles @xmath29 , @xmath30 , @xmath31 and @xmath32 ( radians ) that are separated by a distance of 1.5 mv at 45 degrees that represents the interdot capacitive coupling ( see inset of fig . [ fig : fig3 ] of the main text ) . the width of these lines ensures that experimentally measured crossings still overlap with the template despite unavoidable small variations in the interdot capacitive coupling and the lever arms between gate voltage and electrochemical potential , which affect the slope of the transitions . the final step consists of checking whether extra charging lines are visible in a region of @xmath870@xmath1770 mv@xmath18 to the side of more negative gate voltages . the size of this region should be larger than the charging energy of each dot in mv . the topright corner of the 70@xmath1770 mv@xmath18 area is located -10 mv southwest of the most bottom - left cross . we slightly extend this region on top and on the right to reduce the probability that a charging line is missed . if the total region falls outside the scan range of the data , the algorithm reduces the size of this region accordingly ( alternatively , one could take data over a larger gate voltage range ) . the algorithm could then draw the wrong conclusion . when the region that results from clipping at the border of the scan range is smaller than 40@xmath1740 mv@xmath18 , the algorithm will stop and output that it can not properly determine whether the single - electron regime has been attained . in the typical case that the region is large enough , we first smoothen the data within this region . we subtract the smoothed data from the original data and check whether the resulting pixel values fall above a certain threshold that is proportional to the standard deviation of the smoothed dataset . if at most one pixel value is larger than the threshold , the algorithm classifies the dataset as ` single - electron regime ' . 4 ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1 '' '' @noop [ 0]secondoftwo sanitize@url [ 0 ] + 12$12 & 12#1212_12%12 @startlink[1 ] @endlink[0 ] @bib@innerbibempty http://www.python.org [ `` , '' ] http://www.scipy.org/ [ `` , '' ] ( ) , \url{https://en.wikipedia.org / w / index.php?title = gabor_filter&oldid=681006471 } [ `` , '' ] ( ) , link:\doibase 10.1364/josaa.2.001160 [ * * , ( ) ]	we report the computer - automated tuning of gate - defined semiconductor double quantum dots in gaas heterostructures . we benchmark the algorithm by creating three double quantum dots inside a linear array of four quantum dots . the algorithm sets the correct gate voltages for all the gates to tune the double quantum dots into the single - electron regime . the algorithm only requires ( 1 ) prior knowledge of the gate design and ( 2 ) the pinch - off value of the single gate @xmath0 that is shared by all the quantum dots . this work significantly alleviates the user effort required to tune multiple quantum dot devices . electrostatically defined semiconductor quantum dots have been the focus of intense research for the application of solid - state quantum computing @xcite . in this architecture , quantum bits ( qubits ) can be defined by the spin state of an electron . recently , several experiments have shown coherent manipulation of such spins for the purpose of spin - based quantum computation @xcite . enabled by advances in device technology , the number of quantum dots that can be accessed is quickly increasing from very few to many @xcite . up to date , all these quantum dots have been tuned by ` hand ' . this is a slow process whereby gate voltages are tweaked carefully , first to reach a regime with one electron in each of the dots , and then to adjust the strength of all the tunnel barriers . defects and variations in the local composition of the heterostructure lead to a disordered background potential landscape , which must be compensated for by the gate voltages . on top , cross - capacitances of each gate to neighboring dots increases the tuning complexity as the number of dots increases . the ability to tune these dots automated by computer algorithms , including tuning of many dots in parallel , is an important ingredient towards the scalability of this approach to create a large - scale quantum computer . + in this letter , we demonstrate the computer automated tuning of double quantum dot ( dqd ) devices . we have created an algorithm that only requires as input : ( 1 ) prior knowledge of the gate design , which is reasonable for future large - scale quantum dot circuits and ( 2 ) the measured pinch - off value of the single gate @xmath0 shared by all the quantum dots . we describe the algorithm used and verify its robustness by creating three independent dqds inside a quadruple dot array . the algorithm finds the correct gate voltages to tune all dqds into the single - electron regime and the computer recognizes that this goal has been achieved within an overnight measurement . + a scanning electron microscopy ( sem ) image of a device nominally identical to the one used is shown in fig . [ fig : fig1](a ) . gate electrodes fabricated on the surface of a gaas / algaas heterostructure are biased with appropriate voltages to selectively deplete regions of the two - dimensional electron gas ( 2deg ) 90 nm below the surface and define the quantum dots . the main function of each gate is as follows : gates @xmath1 and @xmath2 set the tunnel coupling with the left and right reservoir , respectively . @xmath3 control the three inter - dot tunnel couplings and @xmath4 are used to set the electron number in each dot . however , each gate influences the other parameters as well . changing @xmath1 for example , will also change the electron number in dot 1 and influence the inter - dot tunnel barrier between dot 1 and 2 . this needs to be taken into account by the algorithm . two other nearby quantum dots on top of the qubit array , sensing dot 1 and 2 ( sd1 and sd2 ) , are created in a similar way and function as a capacitively coupled charge sensor of the dot array . when positioned on the flank of a coulomb peak , the conductance through the sensing dot is very sensitive to the number of charges in each of the dots in the array . changes in conductance are measured using radiofrequency ( rf ) reflectometry @xcite . high - frequency lines are connected via bias - tees to gates @xmath5 , @xmath6 and @xmath7 . the device was cooled inside a dilution refrigerator to a base temperature of @xmath815 mk . all measurements were taken at zero magnetic field . + before running the algorithm the user is required to input a range of @xmath0-values for which the algorithm should try to find dqds . this range is currently determined by measuring the pinch - off value of @xmath0 manually , and then choosing a set of gate voltages more negative than this pinch - off value . the pinch - off value can for example be determined by setting all other gates to 0 mv and next measuring the current from o1 to o4 ( other ohmics open ) whilst sweeping @xmath0 . this step could be automated in future work . + the algorithm consists of 3 steps : ( 1 ) to determine the starting values for the gate voltages , we first measure the pinch - off characteristic between each individual gate and the shared @xmath0-gate . based on those results we ( 2 ) create single quantum dots . the required tunnel barriers acquired in ( 2 ) can be used as a starting point to ( 3 ) create double dots into the single - electron regime . subsequently , steps ( 1 ) and ( 2 ) are used to create the sds . + to measure the pinch - off characteristic we apply a small voltage - bias ( @xmath9v ) to o4 and measure the current @xmath10 through the quadruple dot array . variations in the local composition of the heterostructure underneath each gate will be reflected in the required voltage to create quantum point contacts ( qpcs ) . we term this voltage the transition value , @xmath11 , which is defined as the gate voltage for which @xmath10 is at @xmath12% of its maximum value ( see supplementary information [ sec : pinch_off ] ) . this procedure is repeated for a range of @xmath0-values . figs . [ fig : fig1](b - d ) show an example for @xmath13 mv and the gates controlling the leftmost dot ( @xmath1 , @xmath5 and @xmath14 ) . in practice , it is best to continue with the most positive @xmath0-value that still allows pinch - off for all gates . in our experience this tends to create better quantum dots for this gate design . + , is measured to create single dots . the reflectance of the sds , @xmath15 and @xmath16 , is monitored to tune dqds into the single - electron regime . ( b - d ) pinch - off curves showing @xmath10 versus the gate voltages @xmath1 , @xmath5 and @xmath14 ( @xmath13 mv ) . the calculated transition value is shown as a dotted vertical line . the horizontal lines denote the high and low value as detected by the algorithm ( see supplementary information [ sec : pinch_off]).,scaledwidth=50.0% ] we start by creating single quantum dots , as they already include much of the cross - talk between gates , dots and barriers , that is present in double dots . to create single quantum dots we apply a fixed voltage for the plunger gate ( usually -80 mv ) which we know is appropriate for this device design , and use the transition values of the barrier gates as input for a 2d coarse scan . a suitable scan range is [ @xmath11 - 10 mv , @xmath11 + 400 mv ] . we again monitor @xmath10 . the structure of these scans is always similar : for negative gate voltages the channel is closed , so there is no current . for more positive voltages the channel is open , so there is a large current . we fit a tetragon to the area corresponding to large current , see fig . [ fig : refinement](a ) for an example of the leftmost dot ( details can be found in supplementary information [ sec : analysis_of_single_dots ] ) . we next take a finer scan of the area closest to the tetragon corner with the most negative gate voltages , see figs . [ fig : refinement](b - e ) . in the experiments we have performed , this point is always showing the start of quantum dot formation through the appearance of a coulomb peak . we use this point as the starting point in gate - space for creating dqds . the exact location of the coulomb peak is determined using a gabor filter and is shown as black dots in figs . [ fig : refinement](b - e ) ( see supplementary information [ sec : analysis_of_single_dots ] ) . + when going to double dots , transport measurements are not suitable as current levels through few - electron double dots are impractically low for this device design . therefore , once the single dots have been formed , we tune the sds in a similar way . they can then be used for non - invasive charge sensing which does allow one to distinguish single - electron transitions in the dot array through rf - reflectometry . to achieve a high sensitivity it is important that the sd is tuned to the flank of one of its coulomb peaks . after finding a coulomb peak for the sd in a similar way as described for the qubit dots , we make a 1d scan of the plunger gates , see fig . [ fig : refinement](f ) . each detected coulomb peak is given a score based on its height and slope that allows the algorithm to continue with the most sensitive operating point for the corresponding plunger gate ( see supplementary information [ section : coulumbtuning ] ) . + with the sd tuned we create a double dot in the following way : first we set the voltages of the gates for the double dot to the values found for the individual single dots ( black dots in figs . [ fig : refinement](b - e ) ) . for the single gate shared by the two individual dots ( e.g. gate @xmath14 for the leftmost double dot ) the average of the two values is used . next , we record a charge stability diagram of the double dot structure by varying the two plunger gate voltages involved . we use a heuristic formula to determine the correct scan range that takes into account the capacitive coupling of the gates to the dots ( see supplementary information [ sec : tuning_double_dot ] ) . typical results for such scans are shown in fig . [ fig : fig3](a - c ) . scans involving two plungers are measured by applying a triangular voltage ramp to the plungers on the horizontal axis using an arbitrary waveform generator , and by stepping the other plunger gate using dacs @xcite . whilst stepping the latter gate we also adjust the sensing dot plunger gate to compensate for cross - capacitive coupling and thereby improve the operating range of the sd . + to verify that the double dot has reached the single - electron regime , the algorithm first detects how well specific parts of the charge stability diagrams match the shape of a reference cross ( see inset of fig . [ fig : fig3 ] ) . each match should ideally correspond to the crossing of a charging line from each dot . the shape of the reference cross is derived from the various capacitive couplings , which follow from the gate design and are known approximately from the start . instead of detecting crosses , one could also try to detect the individual charge - transition lines . this turned out to be more sensitive to errors for two reasons : ( 1 ) extra features in the charge stability diagrams that do not correspond to charging lines are wrongfully interpreted as dot features . ( 2 ) not all charging lines are straight across the entire dataset ; this makes it harder to interpret which line belongs to which dot . the cross - matching algorithm is robust against such anomalies because of the local , instead of global , search across the dataset . in future work it could actually be useful to still detect these extra and/or curved lines . they could give information about e.g. unwanted additional dots and aid in determining the electron numbers in regions with higher tunnel couplings . for the current goal of finding the single - electron regime this extra information is not required . + next , the algorithm checks whether within a region slightly larger than 70@xmath1770 mv@xmath18 , it finds other charge transitions for more negative gate voltages with respect to the most bottom - left detected cross ( see supplementary information [ sec : tuning_double_dot ] ) . these regions are depicted by the green tetragons in fig . [ fig : fig3 ] . if no extra transitions are detected : the single - electron regime has been found and the result is given a score of 1 for that specific measurement outcome . if extra transitions are found the algorithm outputs the score 0 . in both cases this is where the algorithm stops . at the end of the run the user can see the measurement results for the various initial choices of @xmath0 and select the best one . + all combined , the running of this complete algorithm ( for a single value of the @xmath0-gate ) takes @xmath19 minutes . per device typically 5 @xmath0-values are tested . in practice we have observed that for some cooldowns of the sample the algorithm could not attain the single - electron regime . a thermal cycle combined with different bias cooling @xcite can significantly influence the tuning and solve this issue ; just as for tuning done by hand . the key difference is that with the computer - aided tuning hardly any user effort is required to explore tuning of double dots to the few - electron regime . in future work the time required for automated tuning ( as well as for tuning by hand ) can be further reduced by also connecting the tunnel barrier gates of each single dot to a high - frequency line which would allow much faster scans for figs . [ fig : fig1]-[fig : refinement ] @xcite . these scans currently form the bottleneck in the overall tuning process . future experiments will also address the automated tuning of more than two dots and the tuning of the tunnel couplings in between dots and their reservoirs , which are key parameters for operating dots as qubit devices . + in summary , we have demonstrated computer - automated tuning of double quantum dot devices into the single - electron regime . this work will simplify tuning dots in the future and forms the first step towards automated tuning of large arrays of quantum dots . the authors acknowledge useful discussions with the members of the delft spin qubit team , and experimental assistance from m. ammerlaan , j. haanstra , r. roeleveld , r. schouten , m. tiggelman and r. vermeulen . this work is supported by the netherlands organization of scientific research ( nwo ) graduate program , the intelligence advanced research projects activity ( iarpa ) multi - qubit coherent operations ( mqco ) program and the swiss national science foundation . 14 ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1 '' '' @noop [ 0]secondoftwo sanitize@url [ 0 ] + 12$12 & 12#1212_12%12 @startlink[1 ] @endlink[0 ] @bib@innerbibempty http://link.aps.org/doi/10.1103/revmodphys.79.1217 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.85.961 [ * * , ( ) ] link:\doibase 10.1146/annurev - conmatphys-030212 - 184248 [ * * , ( ) ] link:\doibase 10.1126/science.1116955 [ * * , ( ) ] link:\doibase 1148092 [ pii ] 10.1126/science.1148092 [ * * , ( ) ] link:\doibase 10.1038/nnano.2013.168 [ * * , ( ) ] link:\doibase 10.1038/nnano.2014.153 [ * * , ( ) ] link:\doibase 10.1038/nnano.2014.216 [ * * , ( ) ] @noop * * ( ) @noop * * ( ) link:\doibase 10.1103/physrevb.81.161308 [ * * , ( ) ] http://arxiv.org/abs/1507.07991 [ ( ) ] link:\doibase 10.1016/j.physe.2006.03.118 [ * * , ( ) ] link:\doibase 10.1103/physrevapplied.4.014018 [ * * , ( ) ] * supplementary material *
in iron - based superconductors ( fescs ) , it is very important to understand the electron correlation in different bands and the orbital selective mott transitions . as revealed by angle resolved photoemission spectroscopy ( arpes ) and band structure calculations , the superconductivity and normal state of fescs are governed by their electronic structure involving the fe 3d orbitals crossing the fermi energy@xcite . due to different structures and probably different methods for fabrication , the iron chalcogenide superconductors show up with different superconducting transition temperatures , for example , @xmath3 8 k for fe@xmath4se single crystals@xcite , @xmath3 16 k for fese@xmath4te@xmath0@xcite , @xmath3 32 k for k@xmath0fe@xmath1se@xmath2@xcite . furthermore , a superconducting - like gaped feature can even be found in the monolayer thin film of fese / srtio@xmath5 at 65 k@xcite , this brings about new vitality in exploring high temperature superconductivity in the iron pnictide / chalcogenide systems . in 2010 , the discovery of the new k@xmath0fe@xmath1se@xmath2 superconductor has generated great excitement in the community@xcite . previously , it was proposed that the pair - scattering of electrons between the hole and electron pockets might drive the electron pairing with an @xmath6 pairing symmetry@xcite . nevertheless , the superconducting k@xmath0fe@xmath1se@xmath2 seems containing only electron pockets around the m - point , the hole pockets which appear in most feas based superconductors near the @xmath7-point are absent@xcite , which certainly leads to a challenge to explain the superconductivity by the nesting effect between the hole and electron pockets . later on , it is found that the k@xmath0fe@xmath1se@xmath2 sample separates into two phases@xcite - a dominant antiferromagnetic insulating phase k@xmath2fe@xmath8se@xmath9@xcite , and a minority superconducting phase whose exact structure and compositions are still under debate@xcite . for the superconducting phase , it has been gradually conceived that the main structure is still the same as the typical bafe@xmath2as@xmath2 , but both the fe and k may be deficient , and these deficiencies may form some kind of structures . recently , the neutron diffraction experiment indicates a potassium deficient but iron stoichiometric formula k@xmath0fe@xmath2se@xmath2 for the superconducting phase@xcite . therefore the study of such material is complicated by the nature of mesoscopic phase separation@xcite , and it is very challenging to explore the properties of the normal state of the superconducting phase . recently , yi _ et al_. reported the study of arpes on a@xmath0fe@xmath1se@xmath2 ( a = k , rb ) single crystals , and proposed an orbital - selective mott transition in the normal state@xcite . moreover , femtosecond pump - probe spectroscopy@xcite and thz spectroscopy@xcite studies also evidenced the mott - transition - like behavior in the normal state . these properties are consistent with the prediction of a multiband theory assuming strong on - site coulomb interactions@xcite . due to the reactivity of potassium element , the sample of k@xmath0fe@xmath1se@xmath2 is very sensitive in air , therefore , as far as we know , the hall effect has been seldom investigated in k@xmath0fe@xmath1se@xmath2 . previous preliminary measurements reveal that the hall coefficient is negative over the whole temperature range , indicating that the system is dominated by electronic - like charge carriers@xcite , being consistent with the observation of the electron fermi pockets in the arpes measurements . in this paper , we study the normal - state transport properties of quenched k@xmath0fe@xmath1se@xmath2 superconducting single crystals with @xmath10 = 32 k through the in - plane resistivity , transverse magnetoresistance ( mr ) and hall effect measurements . we find that the kohler s rule is obeyed in the low temperature region . the hall coefficient has a strong but non - monotonic temperature dependence below 150 k. this is in contrast with the feas - based systems in which the hall coefficient shows a monotonic temperature dependence@xcite . these abnormal temperature - dependent behaviors can not be described by one single band model , suggesting the multi - band nature in k@xmath0fe@xmath1se@xmath2 . using a two - band model ( mainly @xmath11 and @xmath12 ) analysis , we conclude that a gap may open below 65 k , while the non - monotonic temperature dependence of the hall coefficient could be understood as a consequence of an orbital - selective mott transition . these results would trigger further theoretical and experimental studies of the orbital selective correlation effect in fescs . fe@xmath1se@xmath2 single crystal with a broad hump appearing at @xmath13 240 k. a sharp superconducting transition is observed at about 32 k. an enlarged view near the superconducting transition is shown as an inset . , width=340 ] fe@xmath1se@xmath2 . ( a ) the mr below 80 k. solid lines are the fits as a function of @xmath14 . the inset shows temperature dependence of mr measured at a magnetic field of 5 t. a pronounced minimum occurs at 80 k. the red line is a fit to the data below 80 k as a function of @xmath15 . ( b ) the mr above 85 k plotted verse @xmath14.,width=340 ] the k@xmath0fe@xmath1se@xmath2 single crystals used for the transport measurements were fabricated by a self - flux method with a starting synthesizing composition of k : fe : se=0.8:2:2 . the crystals are rapidly quenched in liquid nitrogen after heating to 350 @xmath16c and staying for several hours . details of the preparation were described elsewhere@xcite . the quenching process can greatly improve the connection of the tiny superconducting networks ( paths ) , and thus the global appearance of superconductivity is much better than the slowly cooled samples of k@xmath0fe@xmath1se@xmath2@xcite . x - ray diffraction patterns ( xrd ) taken on the quenched crystals show only ( 00l ) peaks with some small accompanying peaks . we think that the main diffraction ( 00l ) peaks of the xrd pattern are coming from the major part of the sample , i.e. , the k@xmath2fe@xmath8se@xmath9 matrix , while the small accompanying peaks are coming from the minority superconducting networks . because there is clear evidence of phase separation in the sample , it is meaningless to claim an uniform composition through out the sample . the microanalysis using energy - dispersive - spectrum ( eds ) on the samples reveals that the background has a composition close to k : fe : se=2:4:5 . for the present quenched sample , since the minority superconducting phase ( path ) has very small size which is smaller than the size of the electron beam in the eds analysis , we could not use the eds technique directly to get valid values of compositions for the three elements@xcite . transport measurements were carried out with the six - lead method in a quantum design instrument physical property measurement system ( ppms ) . the electric contacts were made using silver paste in a glove box filled with nitrogen atmosphere . we have worked on two samples from different batches and the results are similar to each other . in the k@xmath0fe@xmath1se@xmath2 system , one concern is the phase - separation property , the superconducting and insulating phases could both contribute in transport measurements . as shown in fig . 1 , a broad hump with the peak at @xmath13 240 k is observed in the normal - state , being similar to those reported previously@xcite . this anomaly , being sensitive to the preparation process , could be caused by the connection in series between the metallic and the insulating phases as a result of the phase - separation picture . hence , we mainly focus on the normal - state properties below @xmath17 which are dominated by contributions from the metallic phase ( superconducting paths below @xmath10 ) . the small residual resistivity with sharp superconducting transition ( @xmath10 = 32 k ) indicates good connectivity of the superconducting paths . in order to get more information of the normal - state properties , further investigations are provided by the transverse mr and hall effect measurements on the same sample . in general , for most normal metals , the mr exhibits a @xmath14-dependence in the weak - field limit , and the mr normally affords a useful method to investigate the nature of electronic scattering . in fig . 2(a ) , we show the field dependence of the transverse mr , @xmath18 , below 80 k , solid lines are fits to @xmath19 , @xmath20 is the resistivity under zero magnetic field . it is clear that the mr increases as @xmath14 in the sweeping magnetic field up to 5 t , signaling metallic behavior for _ t _ @xmath21 80 k , while the slight deviation at 35 k can be attributed to the superconducting fluctuations at a finite magnetic field . besides , there is a fundamental difference in the mr between _ @xmath21 80 k and t @xmath22 85 k. in the latter , as shown in fig . 2(b ) , the mr seems not follow the @xmath14-dependence , which indicates a transition of the electronic characteristics from one to another . the inset of fig . 2(a ) shows the temperature dependence of mr at 5 t , a clear minimum occurs at about 80 k and the mr below 80 k shows a nice fit to the @xmath15-dependence . it is interesting to note that , the crossover at around 80 k is consistent with the measurement of the finite - frequency dielectric function by means of terahertz spectroscopy in a rb - based sister compound@xcite . the kohler s rule@xcite , which assumes a simple scaling function of @xmath23 , should be satisfied for a single band metal with an isotropic fermi surface , with @xmath20 the resistivity at a fixed temperature and zero field . for a multiband system , this rule is also applicable as long as the number of charge carriers from each band is independent on temperature and the scattering rates of different bands have the similar temperature dependence@xcite . from the first glance at the data below 60 k , as shown in fig . 3(a ) , it seems that the kohler s rule is slightly violated . actually this slight `` deviation '' of kohler s rule may not be true . the reason is that the resistivity at zero field , namely @xmath24=@xmath25 does not really reflect directly the scattering rate @xmath26 when the residual resistivity is sizable . in addition , as argued below , there maybe a partial gap openning below 65 k , which may give an influence of the kohler s scaling rule . instead of using the original scaling function @xmath27 , we use here a more accurate form of the kohler s scaling rule @xmath28 where @xmath29 is the relaxation rate@xcite . since the system exhibits metallic properties in the low temperature region , we could assume the relaxation rate as @xmath30 . 3(b ) shows the refined kohler s plot of @xmath31 vs @xmath32 . one can see that the kohler s rule is well obeyed below 80 k if we assume a general scattering rate @xmath33 . in contrast , the kohler s rule is drastically violated above 85 k , as shown with the enlarged view in the inset of fig . we now switch our attention to the temperature dependence of hall coefficient . in the present k@xmath0fe@xmath1se@xmath2 system , hall effect measurements may provide the message concerning the temperature dependence of the charge carrier density and mobilities of electrons in different bands of the superconducting phase . since the insulating phase k@xmath2fe@xmath8se@xmath9 has the nearest band 300 mev below the fermi energy , they should not contribute in the hall effect measurements below t@xmath34 . in fig . 4(a ) , we show the hall resistivity @xmath35 versus magnetic field up to 5 t , a linear relation between @xmath35 and magnetic field @xmath36 has been found in wide temperature region ( 35 k to 150 k ) . from the @xmath37 data , the hall coefficient @xmath38 is determined through @xmath39 and shown in fig . the negative @xmath38 over the whole temperature region up to 150 k reveals that the conduction is dominated by electron - like charge carriers . however , the most remarkable feature in fig . 4(b ) is that the @xmath40 shows a strong but non - monotonic temperature dependence . this is in sharp contrast with the feas - based 122 samples in which the hall coefficient is monotonically dependent on temperature@xcite . by having a closer scrutiny to the temperature dependence of hall coefficient , two characteristic temperatures could be defined : @xmath41 = 65 k and @xmath42 = 85 k. below 65 k , @xmath38 decreases rapidly with a suppression towards lower temperatures . this is quite similar to that in the feas - based superconductors . between 65 k and 85 k , @xmath38 is almost temperature independent , while it decreases upon raising temperature from 85 k to 150 k. this anomalous behavior has never been reported in previous studies in k@xmath0fe@xmath1se@xmath2 and suggests that something beyond the multi - band physics is very important here in determining the electric conduction . at various denoted temperatures . inset shows the kohler s scaling of mr in low field region for selective temperatures . ( b ) a refined kohler s plot below 80 k using @xmath28 and assuming @xmath43.,width=340 ] , two characteristic temperatures could be defined : @xmath41 = 65 k and @xmath42 = 85 k. solid lines are theoretical curves using the two - band model ( see text).,width=340 ] according to the band structure calculations and arpes studies , two sets of electronic orbitals near the fermi level , namely @xmath12 and @xmath11 play an important role . since the @xmath44 and @xmath45 are normally degenerate , we thus use a two band model to handle the issue . based on the boltzmann transport theory in the weak field limit , the equation of the hall coefficient for two - band model could be simplified as @xmath46 where @xmath47 and @xmath48 are single band conductivity and hall coefficients for the two orbitals , respectively . based on the mr analysis , the system exhibits metallic state below 80 k and a @xmath15 temperature dependence was found in our data . thus , the mobility of the two orbitals are expressed generally as @xmath49 , @xmath50 is a parameter related to the effective density of states . since the kohler s rule is well obeyed below 80 k , concerning the fact that the recent thz spectroscopy experiments on rb@xmath4fe@xmath1se@xmath2 report a gap - like suppression of optical conductance below 61 k@xcite , we thus attribute the temperature dependence of the hall coefficient below 65 k to the temperature dependent of the charge carrier density . in this scenario we then define the carrier densities as @xmath51 and @xmath52 where @xmath53 is a partial gap which opens at 65 k. the red line in fig . 4(b ) presents excellent fitting results with the two - band model . one may suggest that the temperature dependence of the hall coefficient @xmath38 below 65 k is related to the different temperature dependence of the scattering rate @xmath54 , this however can not get support from other experiments , for example the arpes and optical data . after crossing a maximum , @xmath38 drops with increasing temperature above 85 k , which signals the involvement of a distinct electron scattering mechanism . this turning point and the change of the temperature dependence is consistent with the results of mr analysis . recently , an orbital - selective mott transition was observed by arpes at about 90 k in the superconducting phase of k@xmath0fe@xmath1se@xmath2 , where the spectral weight near fermi surface for the @xmath12 orbital diminishes while the other orbitals @xmath11 remain metallic@xcite . moreover , investigations using pump - probe spectroscopy@xcite and thz spectroscopy@xcite also evidenced the mott - transition related behavior in the normal state . this scenario could give a reasonable explanation to our data here . since the @xmath11 orbital remains metallic , we still express the mobility of @xmath11 orbital as @xmath55 and the carrier density as a constant @xmath56 . meanwhile , we consider that the @xmath12 band goes into the mott phase with raising temperature . as it is well known , the mott insulating behavior has been experimentally identified and theoretically explained in terms of the band narrowing effect associated with the electron - electron correlation . therefore , the mobility of @xmath12 orbital could be interpolated with the formula @xmath57 , where @xmath58 is the modification term associated with the mott transition . to approach a solution , we may set the carrier density of the @xmath12 orbital as a constant @xmath59 for simplicity . thus , the expression of the hall coefficient with the orbital - selective mott phase is written as @xmath60 where @xmath61 is the relative ratio of the mobility coefficient of the two orbitals . in fig . 4(b ) , the blue solid line above 85 k shows the theoretical fitting result . consequently , we acquire the parameters as @xmath53 = 7.7 mev , @xmath62 m@xmath63 , @xmath64 m@xmath63 , @xmath65 m@xmath63 , @xmath66 and @xmath67 = 0.001 k@xmath68 . the plateau of the hall coefficient r@xmath34 between 65 k and 85 k may be viewed as the crossover of the two different regions . we note that the appearance of a gap with the value of @xmath53 = 7.7 mev below 65 k could be compared to the observation of high - temperature superconductivity at 65 k in single - layer fese films@xcite . we also ca nt rule out the possibility of a pseudogap opening or other explanations for this gap - like suppression of @xmath38 below 65 k. the decrease of @xmath38 for increasing temperature starting from 85 k can get a strong support from the scenario of the orbital - selective mott transition in this system . the explanation based on the orbital selective mott transition should call for further theoretical and experimental efforts . in summary , magnetoresistance and hall coefficient @xmath38 have been measured in superconducting k@xmath0fe@xmath1se@xmath2 single crystals . the kohler s rule is well obeyed below 80 k by assuming a general scattering rate @xmath69 . we have observed a strong and non - monotonic temperature dependence of the hall coefficient in the normal state . using a two - band model analysis and combining with the published data of the time domain optical conductivity measurements , we conclude that a gap may open below 65 k , while the data above 85 k could be understood as a consequence of an orbital - selective mott transition of the @xmath12 band . we appreciate the useful discussions with qimiao si and the help of baigen wang . this work is supported by the nsf of china ( 11034011/a0402 ) , the ministry of science and technology of china ( 973 projects : 2011cba00102 and 2012cb821403 ) and papd . p. hirschfeld , m. korshunov , and i. mazin , reports on progress in physics * 74 * , 124508 ( 2011 ) . hsu , j .- y . luo , k .- w . yeh , t .- k . chen , t .- w . huang , p. m. wu , y .- c . lee , y .- l . huang , y .- y . yan , et al . , proceedings of the national academy of sciences * 105 * , 14262 ( 2008 ) . yeh , t .- w . huang , y .- l . huang , t .- k . chen , f .- hsu , p. m. wu , y .- c . lee , y .- y . chen , j .- y . luo , et al . , epl * 84 * , 37002 ( 2008 ) . j. guo , s. jin , g. wang , s. wang , k. zhu , t. zhou , m. he , and x. chen , phys . b * 82 * , 180520 ( 2010 ) . m fang , h .- wang , c .- h . dong , z .- j . li , c .- m . feng , j. chen , and h. yuan , epl * 94 * , 27009 ( 2011 ) . d. liu , w. zhang , d. mou , j. he , y .- b . ou , q .- y . wang , z. li , l. wang , l. zhao , s. he , et al . , nature communications * 3 * , 931 ( 2012 ) . s. he , j. he , w. zhang , l. zhao , d. liu , x. liu , d. mou , y .- b . ou , q .- y . wang , z. li , et al . , nature materials * 12 * , 605 ( 2013 ) . h .- wen , reports on progress in physics * 75 * , 112501 ( 2012 ) . i. mazin , d. j. singh , m. johannes , and m .- h . du , phys . * 101 * , 057003 ( 2008 ) . k. kuroki , s. onari , r. arita , h. usui , y. tanaka , h. kon- tani , and h. aoki , phys . * 101 * , 087004 ( 2008 ) . i. shein and a. ivanovskii , phys . a * 375 * , 1028 ( 2011 ) . b. shen , b. zeng , g. chen , j. he , d. wang , h. yang , and h. wen , epl * 96 * , 37010 ( 2011 ) . w. li , h. ding , p. deng , k. chang , c. song , k. he , l. wang , x. ma , j .- p . hu , x. chen , et al . , nature physics * 8 * , 126 ( 2012 ) . y. zhang , l. yang , m. xu , z. ye , f. chen , c. he , h. xu , j. jiang , b. xie , j. ying , et al . , nature materials * 10 * , 273 ( 2011 ) . d. ryan , w. rowan - weetaluktuk , j. cadogan , r. hu , w. straszheim , s. budko , and p. canfield , phys . b * 83 * , 104526 ( 2011 ) . w. bao , q .- z . huang , chen , m. a. green , d .- wang , j .- b . he , and y .- m . * 28 * , 086104 ( 2011 ) . f. ye , s. chi , w. bao , x. wang , j. ying , x. chen , h. wang , c. dong , and m. fang , phys . lett . * 107 * , 137003 ( 2011 ) . w. li , h. ding , z. li , p. deng , k. chang , k. he , s. ji , l. wang , x. ma , j .- hu , et al . , phys . * 109 * , 057003 ( 2012 ) . y. texier , j. deisenhofer , v. tsurkan , a. loidl , d. inosov , g. friemel , and j. bobroff , phys . 108 * , 237002 ( 2012 ) . x. ding , d. fang , z. wang , h. yang , j. liu , q. deng , g. ma , c. meng , y. hu , and h .- h . wen , nature communications * 4 * , 1897 ( 2013 ) . s. v. carr , d. louca , j. siewenie , q. huang , a. wang , x. chen , and p. dai , phys . b * 89 * , 134509 ( 2014 ) . m. yi , d. lu , r. yu , s. riggs , j .- h . chu , b. lv , z. liu , m. lu , y .- t . cui , m. hashimoto , et al . lett . * 110 * , 067003 ( 2013 ) . w. li , c. zhang , s. liu , x. ding , x. wu , x. wang , h .- h . wen , and m. xiao , phys . b * 89 * , 134515 ( 2014 ) . z. wang , m. schmidt , j. fischer , v. tsurkan , m. greger , d. vollhardt , a. loidl , and j. deisenhofer , nature com- munications * 5 * ( 2014 ) . r. yu and q. si , phys . rev * 110 * , 146402 ( 2013 ) . x. luo , x. wang , j. ying , y. yan , z. li , m. zhang , a. wang , p. cheng , z. xiang , g. ye , et al . , new j. phys.13 , 053011 ( 2011 ) . p. cheng , h. yang , y. jia , l. fang , x. zhu , g. mu , and h .- h . wen , phys . b * 78 * , 134508 ( 2008 ) . h. luo , p. cheng , z. wang , h. yang , y. jia , l. fang , c. ren , l. shan , and h .- h . wen , physica c : superconductivity * 469 * , 477 ( 2009 ) . l. fang , h. luo , p. cheng , z. wang , y. jia , g. mu , b. shen , i. mazin , l. shan , c. ren , et al . b * 80 * , 140508 ( 2009 ) . f. han , h. yang , b. shen , z .- y . wang , c .- h . li , and h .- h . wen , philosophical magazine * 92 * , 2553 ( 2012 ) . y. liu , q. xing , k. w. dennis , r. w. mccallum , and t. a. lograsso , phys . b * 86 * , 144507 ( 2012 ) . a. charnukha , j. deisenhofer , d. prpper , m. schmidt , z. wang , y. goncharov , a. n. yaresko , v. tsurkan , b. keimer , a. loidl , and a. v. boris , phys . b * 85 * , 100504 ( r)(2012 ) . a. b. pippard , magnetoresistance in metals , 2 ( cambridge university press , 1989 ) . m. chan , m. veit , c. dorow , y. ge , y. li , w. tabis , y. tang , x. zhao , n. barii , and m. greven , arxiv:1402.4472 ( 2014 ) . n. luo , and g. h. miley , physica c * 371 * , 259 ( 2002 ) .	in - plane resistivity , magnetoresistance and hall effect measurements have been conducted on quenched k@xmath0fe@xmath1se@xmath2 single crystals in order to analysis the normal - state transport properties . it is found that the kohler s rule is well obeyed below about 80 k , but clearly violated above 80 k. measurements of the hall coefficient reveal a strong but non - monotonic temperature dependence with a maximum at about 80 k , in contrast to any other feas - based superconductors . with the two - band model analysis on the hall coefficient , we conclude that a gap may open below 65 k. the data above 65 k are interpreted as a temperature induced crossover from a metallic state at a low temperature to an orbital - selective mott phase at a high temperature . this is consistent with the recent data of angle resolved photoemission spectroscopy . these results call for a refined theoretical understanding , especially when the hole pockets are absent or become trivial in k@xmath0fe@xmath1se@xmath2 superconductors .
magnetic activity in the sun is known to play a central role in driving both long - term and short - term dynamics ( tobias 2002 ; weiss 2002 ) . the magnetic field is responsible for spectacular events such as sunspots , solar flares , and coronal mass ejections , and for heating the solar corona to high temperatures . large - scale magnetic activity is known to be dominated by the eleven year activity cycle . this cycle has been systematically observed since the early seventeenth century and its properties are well documented ( see e.g. ossendrijver 2003 ) . of particular current interest is the impact of magnetic activity on solar irradiance that might have significant implications for the terrestrial climate ( see solanki _ et al _ 2004 ) . given the importance of solar activity , it is not surprising that there has been a continued interest in understanding the mechanisms responsible for generating the solar magnetic field . the sun s magnetic field is believed to be generated by a hydromagnetic dynamo in which motion of the solar plasma ( advection ) is able to sustain a magnetic field against the continued action of ohmic dissipation ( see e.g moffatt 1978 ; charbonneau 2005 ) . progress in understanding this fundamental problem of solar magnetohydrodynamics is slow owing to the difficulties of the dynamo problem . the extreme parameters of the solar interior and the inherent three - dimensionality of the dynamo problem make it impossible to solve the equations accurately on a computer . much effort has therefore focused on _ mean - field _ dynamo models ( steenbeck , krause & rdler 1966 ; krause & rdler 1980 ) , which describe the evolution of the mean magnetic field , parameterising the effects of the small - scale fields and flows in terms of tensor transport coefficients . these transport coefficients include @xmath0 ( which leads to a regenerative term in the mean - field equations the so - called @xmath1-effect ) and the turbulent diffusivity ( @xmath2 ) . we stress here that there is no mechanism within the theory for determining the form of these coefficients , except for flows at low magnetic reynolds number or with short correlation time , and in solar models these are usually chosen in a plausible but ad - hoc manner ( often , for simplicity , adopting isotropic representations in which @xmath3 and @xmath4 ) . much attention has been focused upon determining these transport coefficients in both the linear and nonlinear dynamo regimes from numerical simulations ( cattaneo & hughes 1996 ; brandenburg & subramanian 2005 ) but there is still no consensus over the nature of these , even to within an order of magnitude ( see courvoisier , hughes & tobias 2006 ) . mean - field models have , however , proved successful in providing illustrations of the type of behaviour that might be expected to occur in the sun ( and other stars ) . it is often argued that , _ although these models have no predictive power _ , understanding the underlying mathematical form of the equations can lead to the identification of robust patterns of behaviour . many different models have been proposed for the solar dynamo . in the distributed dynamo model , the @xmath1-effect operates throughout the convection zone and interacts with the latitudinal shear ( or the sub - surface shear layer , see brandenburg 2005 ) to generate magnetic field . alternatively , the dynamo could be operating near the tachocline , where an @xmath1-effect might be driven either by a tachocline - based instability or by turbulent convection . this , in conjunction with the strong shear , could drive an `` interface '' dynamo ( parker 1993 ) . finally , there are flux transport models , in which the ( so - called ) babcock - leighton mechanism produces an @xmath1-effect ( or source term ) at the surface . this surface @xmath1-effect is coupled to the radial shear in the tachocline ( where another @xmath1-effect may be operating ) via a meridional flow ( choudhuri , schssler & dikpati 1995 ; dikpati & charbonneau 1999 ) . the relative merits of these models are discussed elsewhere in the literature ( see , e.g. charbonneau 2005 ) the only comment we make here is that this plethora of models arises because of the lack of available constraints on the form of the transport coefficients in the mean - field formalism . we note further that it is not clear that any of the above scenarios capture the essential dynamo processes correctly or that these processes can ever be captured by a mean - field model . it is also possible to construct predictions of solar activity without using dynamo theory , and there is a long literature describing these predictive methods ( see e.g. zhang 1996 ; hathaway , wilson & reichmann 1999 ; sello 2003 ; zhao _ et al _ 2004 ; saba , strong & slater 2005 ) . one class of prediction techniques uses statistical and timeseries analysis methods ( see e.g. tong 1995 for details ) . these methods , which are also applicable in many other areas of physics , vary in complexity from simple linear methods to methods that use dynamical systems theory to reconstruct nonlinear attractors in phase space . however , these methods have the drawback that they do not utilise any of the `` physics '' of the problem . predictions can also be made by using precursor methods ( see e.g. schatten 2002 ) , which do utilise some of the physical features of the system in addition to the timeseries data . in recent papers ( dikpati , de toma & gilman 2006 ; dikpati & gilman 2006 ) , an attempt has been made to unify these two approaches by utilising a mean - field model in order to make predictions about the future activity of the sun . these papers describe an axisymmetric , mean - field model of a flux transport dynamo . here the authors make use of the observations of magnetic flux at the solar surface to feed into a model of solar activity . the flux that is observed at the solar surface is advected by a parameterised meridional flow ( which can be observed down to a certain depth ) and interacts with a differential rotation profile that has been inferred from helioseismology . the magnetic flux also interacts with turbulence , the effects of which are parameterised by certain turbulent transport coefficients ( representing the turbulent diffusivity and the @xmath1-effect ) . as with all current mean field models these turbulent transport effects have been parameterised in a plausible but ad - hoc manner , and are unconstrained by observations and indeed theory . the simplest predictive scheme proposed by dikpati _ et al _ ( 2006 ) therefore takes the form of a parameterised linear system forced by boundary observations . the implicit underlying philosophy here is that by reducing the correct physics for the generation of the solar activity cycle ( i.e. a nonlinear self - excited dynamo ) to such a scheme , predictions about future solar activity can be made . in this paper we shall investigate the predictability of various dynamo models . we demonstrate that even when all the nonlinear physics of the solar dynamo is removed , problems remain for prediction owing to the increased importance of stochastic effects even very weak stochastic perturbations can produce significant modulation in these linear - type models . we also discuss the best - case scenario for prediction where stochastic effects can be ignored , and demonstrate that in these cases prediction is still difficult owing to uncertainties in the input parameters of these parameterised mean - field models . the paper is organised as follows . in the next section we describe ( in a general way ) the importance of modulation and the role of stochasticity and nonlinearity in solar dynamo models . in section 3 we investigate a flux transport model and demonstrate how the presence of even extremely weak noise can render predictions useless . in section 4 we consider the best - case " scenario for prediction where noise does not play a role in the modulation we demonstrate that more accurate prediction schemes may arise by using basic timeseries analysis techniques rather than from constructing mean - field models of the solar cycle . finally , in section 5 we discuss the implications of our work for predictions of the solar cycle . in this section , we discuss the problems that must be overcome by schemes designed to yield a prediction of future solar magnetic activity . some of these problems arise owing to the nature of solar magnetic activity whilst others arise from the lack of a detailed theory that is capable of describing solar magnetic activity in such extreme conditions as those that exist in the solar interior . it is clear that if the solar cycle were strictly periodic , with a constant amplitude , then it would be straightforward to predict future behaviour . however , all measurements of solar magnetic activity ( both direct observations and evidence from proxy data ) indicate that the variations in the magnetic activity do not follow a periodic pattern . departures from periodicity may be driven either by perturbations or by modulation . for the case of a weakly perturbed periodic system , the dynamics is essentially captured by the periodic signal , with the small perturbations playing a secondary role . we distinguish this behaviour from a modulated signal in which there are significant departures from periodicity ( often occurring on longer timescales ) , with large variations in the observed amplitude of the signal . all the evidence from direct observations indicates that the solar cycle is strongly modulated . the amplitude of the solar cycle varies enormously over long timescales , an extreme example of this modulation was a period of severely reduced activity in the seventeenth century known as the maunder minimum . proxy data from records of terrestrial isotopes , such as @xmath5be and @xmath6c ( see e.g. beer 2000 , weiss & tobias 2000 , wagner _ et al _ 2001 ) , demonstrate that this modulation has been a characteristic feature of the solar magnetic activity over ( at least ) the last 20,000 years . mathematically there are only two possible sources for this strong modulation of the basic solar cycle ( tobias 2002 ) . the modulation may arise either as a result of stochastic effects ( see e.g. ossendrijver & hoyng 1996 ) or by deterministic processes ( see e.g. tobias , weiss & kirk 1995 ) . in this context we define deterministic processes to be those that _ are _ captured by the differential equations of dynamo theory , with no random elements . stochastic processes are those that occur on an unresolved length or timescale , and so can not be described by the differential equations without including a random element into the model . it is well known that stochastic modulation can arise even if the deterministic physics that leads to the production of the basic cycle is essentially linear . this parameter regime is generally considered to be a good one for prediction , since any nonlinear effects are only playing a secondary role . however , in this stochastically - perturbed case , the small random fluctuations that lead to the modulation will have large short - term effects and render prediction extremely difficult , if not impossible . conversely , if the modulation arises purely as a result of deterministic processes , then the underlying physics is nonlinear ( or potentially non - autonomous ) and this leads to difficulty in prediction owing to the possible presence of deterministic chaos and ( more importantly ) the difficulty of constructing accurate nonlinear models with large numbers of degrees of freedom . in the next two sections we demonstrate the problems for prediction for dynamo models in both of the classes described above . in the next section we describe a flux transport model of the same type as the one used in the prediction scheme of dikpati _ et al _ ( 2006 ) and we demonstrate that even very small random fluctuations can produce significant modulation , leading to extreme difficulties for prediction . we then , in section 4 , go on to describe a model where the modulation arises owing to the presence of deterministic chaos and show that in this case , prediction using model fitting is a poor way to proceed , but some prediction is possible if it is possible to reconstruct the attractor for activity . we assume initially that the modulated solar magnetic activity can be described by a stochastically - perturbed mean - field dynamo model . in this model , nonlinear effects are playing a secondary role , and all the modulation is being driven by the stochastic effects . the aim of this section is to assess whether or not models of this type can be used to make meaningful predictions of the solar magnetic activity . in these models , the evolution of the large - scale magnetic field is described by the standard mean - field equation ( see , for example , moffatt 1978 ) , @xmath7 here , @xmath8 represents the large - scale magnetic field and @xmath9 corresponds to the mean velocity field , @xmath10 is the ( turbulent ) magnetic diffusivity , and the @xmath11 term corresponds to the mean - field @xmath1-effect . using the well - known @xmath12 approximation , we solve this equation numerically in an axisymmetric spherical shell ( @xmath13 and @xmath14 ) . in solving equation ( [ eqn:1 ] ) we need to ensure that @xmath8 remains solenoidal ( i.e. @xmath15 ) . to achieve this , we decompose the magnetic field into its poloidal and toroidal components , @xmath16 where @xmath17 denotes the toroidal ( azimuthal ) field component and the scalar potential @xmath18 relates to the poloidal component of the magnetic field . so , rather than solving equation ( [ eqn:1 ] ) directly , the problem has been reduced to solving two coupled partial differential equations for the scalar quantities @xmath18 and @xmath17 . we adopt idealised boundary conditions , in which @xmath19 at @xmath20 and @xmath21 and @xmath22 and @xmath23 and @xmath24 are smoothly matched to a potential field at @xmath25 . this particular dynamo model is closely related to the flux transport model described by dikpati & charbonneau ( 1999 ) . the large - scale velocity field , @xmath9 , is given by @xmath26 where @xmath27 is a prescribed analytic fit to the helioseismologically - determined solar rotation profile ( see , for example , bushby 2006 ) and @xmath28 and @xmath29 correspond to a prescribed meridional circulation . we assume that the meridional circulation pattern in each hemisphere consists of a single cell , with a polewards flow at the surface and an ( unobservable ) equatorwards flow at the base of the convection zone the flow is confined to the region @xmath30 . the functional form that we adopt for this flow is similar in form to the one described by dikpati & charbonneau ( 1999 ) , @xmath31 \xi \sin \theta \left ( 3\cos^2 \theta - \sin^2 \theta \right),\label{eqn:4}\\ u_{\theta}(r,\theta)&=&u_o \left ( \frac{r_{\odot}}{r } \right)^3 \left [ -1 + c_1\xi^{0.5}-c_2\xi^{0.75}\right ] \cos \theta \sin^2 \theta,\label{eqn:5 } \end{aligned}\ ] ] where @xmath32 $ ] , @xmath33^{-0.5}$ ] , @xmath34^{-0.75}$ ] , and @xmath35 is some characteristic flow speed . this flow pattern can be stochastically perturbed by setting @xmath36 , where @xmath37 is a time - dependent , randomly fluctuating variable in the range @xmath38 . the aim here is to assess whether or not such weak stochastic variations in the flow pattern could give rise to significant modulation in the activity cycle , and if so what are the consequences for prediction . in order to complete the specification of the model , we need to choose plausible functional forms for the @xmath1-effect and the turbulent magnetic diffusivity . it should be emphasised again that these mean - field coefficients are poorly constrained by theory and observations , although plausible assumptions can be made . defining @xmath39 to be a characteristic value of the turbulent magnetic diffusivity within the solar convection zone , we adopt a similar spherically - symmetric profile to that adopted by dikpati & charbonneau ( 1999 ) , @xmath40 + \beta_c , \label{eqn:6}\ ] ] where erf corresponds to the error function and @xmath41 ( here taken to be @xmath42 of @xmath39 ) represents the magnetic diffusivity below the turbulent convection zone . following dikpati & charbonneau ( 1999 ) , rather than prescribing a simple functional form for @xmath1 we neglect the @xmath1-effect term in the toroidal ( @xmath24 ) field equation and replace the corresponding @xmath43 term in the poloidal ( @xmath23 ) equation by a non - local , nonlinear source of poloidal flux , @xmath44 \left [ 1 - \mbox{erf}\left(\frac{r - r_{\odot}}{0.01r_{\odot } } \right)\right ] \\ \nonumber & & \left [ 1+\left(\frac{b(0.7r_{\odot},\theta , t)}{b_o}\right)^2\right]^{-1}\sin \theta \cos \theta b(0.7r_{\odot},\theta , t).\label{eqn:7}\end{aligned}\ ] ] here , @xmath45 is a characteristic value of this poloidal source and @xmath46 represents the ( somewhat arbitrarily chosen ) field strength at which this non - local source becomes suppressed by the magnetic field . this source term parameterises the contribution to the poloidal magnetic flux due to the decay of active regions the non - locality reflects the fact that active regions are believed to form as the result of buoyant magnetic flux rising from the base of the convection zone to the solar photosphere . see dikpati & charbonneau ( 1999 ) for a more detailed discussion of this source term , though again it must be stressed that the functional form and the nonlinear dependence are chosen in a plausible yet ad - hoc manner . in order to carry out numerical simulations , we first non - dimensionalise this flux transport model . by using scalings similar to those described by dikpati & charbonneau ( 1999 ) , it can be shown that the model solutions are fully determined by two non - dimensional parameters ( once other parameters such as @xmath46 have been selected ) . denoting the equatorial angular velocity at the solar surface by @xmath47 , these non - dimensional parameters are the dynamo number , @xmath48 , and the magnetic reynolds number corresponding to the meridional flow , @xmath49 . here , we set @xmath50 and @xmath51 . in the absence of stochastic noise , this set of parameters produces a strong circulation - dominated dynamo in which the magnetic energy is a periodic function of time . although the dynamo number is not weakly supercritical , nonlinear effects are not strong enough here to produce a modulated activity cycle the primary role of the nonlinearity is to prevent the unstable dynamo mode from growing exponentially . we term such a model a `` linear - type '' model . when weak stochastic effects are included in the model , the resulting activity cycle is indeed weakly modulated . this is illustrated in figure 1 , which shows the time - dependence of this solution . the time - series clearly illustrates that , although the amplitude of the `` cycle minimum '' only appears to be weakly time - dependent , there are significant variations in the peak amplitude of the magnetic energy time - series . these variations are qualitatively similar to those observed by charbonneau & dikpati ( 2000 ) , who considered large amplitude random fluctuations in the flow pattern within the solar convection zone the peak amplitude of these fluctuations was comparable with the peak amplitude of the flow . in this particular model , we have shown that even very weak stochastic variations in the centre of mass of the flow pattern can still produce significantly modulated behaviour . these stochastic effects are expected to become increasingly significant for dynamo numbers approaching critical . so , these models are obviously highly sensitive to the addition of stochastic noise . in the absence of stochastic noise , the attractor ( in phase space ) for this solution is two - dimensional , and the future behaviour of the solution at any instant in time is entirely determined by the current position of the system on the attractor . the same is not true when this system is perturbed by stochastic effects , and it clearly becomes much more difficult to predict the future behaviour of the system . since the attractor of this stochastically perturbed solution can not be unambiguously defined , another possible way of assessing the `` predictability '' of this solution is to look for a correlation between successive cycle maxima . defining @xmath52 to be the magnitude of the @xmath53 cycle maximum , figure 2 shows @xmath54 as a function of @xmath52 . it is clear from this scatter plot that there is no obvious correlation between the amplitudes of successive cycle maxima in this case . since the modulation is being driven entirely by random stochastic forcing , this result is not surprising . this lack of correlation suggests that the behaviour of previous cycles can not be used to infer the magnitude of the following one . this implies that even weak stochastic effects may seriously reduce the possibilities for solar cycle prediction in this linear - type regime . in the previous section , we demonstrated that even very weak stochastic perturbations to the meridional flow pattern can lead to a loss of predictability in a linear - type flux transport dynamo model . in that model , the modulation of the activity cycle was driven entirely by stochastic effects . as discussed in section 2 , the only other possible scenario is that the observed modulation is driven by nonlinear effects . this scenario , where the observed modulation is deterministic in origin , is the `` best - case '' scenario for prediction , as in this case the entirely unpredictable stochastic elements may be ignored . we stress again that , given that solar magnetic activity is significantly modulated , either deterministic or stochastic modulation must be considered in any realistic model ( predictive or otherwise ) of the solar cycle . so , in this section , we completely neglect stochastic effects and assume that the observed ( chaotic ) modulation in the solar magnetic activity can be described by a fully deterministic model in which any activity modulation ( e.g. solar - like `` grand minima '' ) is driven entirely by nonlinear effects . the model that we use was described in detail in two recent papers ( bushby 2005 , 2006 ) , so we only present a brief description here . the exact details of the model are unimportant for our main conclusions . like the flux transport dynamo model from the previous section , this model describes an axisymmetric , mean - field , @xmath12-dynamo in a spherical shell . unlike the previous model , this model represents an `` interface - like '' dynamo that is operating primarily in the region around the base of the solar convection zone . it is worth mentioning again that ( as discussed in the introduction ) there is still no general consensus regarding which of these dynamo scenarios is more likely to be an accurate representation of the solar dynamo . for this interface - like dynamo model , we neglect meridional motions , since they are poorly determined near the base of the solar convection zone . like several earlier models ( e.g. tobias 1997 ; moss & brooke 2000 ; covas _ et al _ 2000 ) , this dynamo model includes the feedback ( via the azimuthal component of the lorentz force ) of the mean magnetic field upon the differential rotation ( malkus & proctor 1975 ) . this nonlinear feedback is a crucial element of the model and , in the absence of stochastic effects , is the sole driver of modulation in the magnetic activity cycle . denoting this magnetically - driven velocity perturbation by @xmath55 , the large - scale velocity field is given by @xmath56\mathbf{e_{\phi } } , \label{eqn:8}\ ] ] where ( as in the previous model ) @xmath27 represents an analytic fit to the solar differential rotation . whilst the evolution of the large - scale magnetic field is again governed by equation ( 1 ) , an additional evolution equation is required for the velocity perturbation , @xmath57 . this equation is given by @xmath58\cdot \mathbf{e_{\phi } } + \frac{1}{r^3}\frac{\partial}{\partial r}\left[\nu r^4 \frac{\partial}{\partial r } \left(\frac{v}{r}\right)\right ] \\ \nonumber & & + \frac{1}{r^2 \sin^2 \theta}\frac{\partial}{\partial \theta}\left[\nu \sin^3 \theta \frac{\partial}{\partial \theta } \left(\frac{v}{\sin \theta}\right)\right],\label{eqn:9}\end{aligned}\ ] ] where @xmath59 represents the fluid density ( here taken to be constant ) , @xmath60 is the permeability of free space and @xmath61 represents the ( turbulent ) fluid viscosity . in order to complete the model , the spatial dependence of the transport coefficients ( @xmath1 , @xmath10 and @xmath61 ) must also be specified . again , we emphasise that there are no direct observational constraints relating to these coefficients as noted in the introduction , there is no consensus as to their form and there is still a debate as to their order of magnitude ( and even their sign ) . having said that , it is possible to make some plausible assumptions for an `` interface - like '' dynamo model ( see bushby 2006 for more details ) . the precise choices of these parameters are unimportant for our main conclusions . having set up this model , it is possible to choose a set of parameters so that the solutions _ do _ reproduce some salient features of the solar dynamo ( bushby 2005 , 2006 ) . we stress here that , although the parameters have been chosen in a plausible manner , this dynamo model should not be regarded as an accurate representation of the solar interior and is subject to many uncertainties . furthermore we stress again that this is the case with _ all _ mean - field solar dynamo models . however we use this model as a useful tool to analyse the possibility of producing predictive models of the solar cycle . we proceed by choosing fiducial parameters and profiles for the turbulent transport coefficients that lead to solar - type " magnetic activity , with chaotically modulated cycles and recurrent `` grand minima '' . we then integrate this model forward in time to produce a timeseries and designate this timeseries as the target " run , which any subsequent model should be able to predict . this target run is shown in figure 3 , which shows a timeseries of the activity together with a reconstruction of the dynamo attractor in phase space . although this solution is chaotically modulated , it is certainly no more chaotic than the equivalent attractor for the @xmath5be data , which is a well - known proxy for solar magnetic activity ( e.g. beer 2000 ) . whilst the nonlinear effects are significant enough to drive the modulation , they are actually very difficult to detect . in this model , the cyclic component of the fluctuations in the differential rotation ( which are driven by the nonlinear lorentz force ) are small compared with the mean differential rotation . this is consistent with observations of the ( so called ) torsional oscillations in the solar convection zone . finally , note once more that , since the modulation is driven entirely by nonlinear effects , this model is specified exactly . the question is then posed as to whether _ any _ mean - field model can be constructed that leads to meaningful predictions of the future behaviour of the target run . clearly the best chance for a mean - field model being capable of predicting the future behaviour of the target run is to use _ the exact model _ that led to the target run data . hence we test this model first , as all subsequent models will be inferior to this . we proceed by setting the model parameters to be those that generated the long test run , and consider the behaviour of solutions that are started from very similar points on the attractor . some of the solutions are shown in figure 4 . this figure shows clearly that although the predictor solutions are able to track the target solution for a couple of activity cycles , the nature of the solutions means that the predictors and target solution diverge quickly after this time . this is not surprising behaviour . it is well - known that chaotic solutions have a sensitive dependence on initial conditions and that long - term prediction of such solutions is fraught with problems ( see e.g. tong 1995 ) . what is clear is that simply using a model that is based upon mean - field theory will not work in the long term _ even if the model is correct in every detail_. one might be able to predict one or two cycles ahead _ if one has solved the problem of constructing an exact representation of the solar dynamo _ but as noted above this is not an easy task . we now turn to the related problem of short - term prediction . as discussed in the introduction there are large uncertainties in the form and amplitude of the input parameters for all mean - field models . what we investigate here is whether these uncertainties lead to significant difficulties in prediction even in the short term . again we examine the best case scenario and consider a mean - field model for the predictor runs that has correctly parameterised the _ form _ of all the input variables ( differential rotation , @xmath1-effect , turbulent diffusivity and nonlinear response ) . in addition these predictors have been given the correct input _ values _ for all - but - one of the parameters . hence the predictor models are exactly the same as the target model with the exception of one input parameter that has been altered by @xmath62 . this would be a staggeringly good representation should it be possible to achieve this for solar activity . furthermore we increase the chances of the predictor being able to predict the future behaviour of the target solution by matching the two timeseries over a number of cycles . this is analogous to the procedure employed by dikpati _ et al _ ( 2006 ) who cite support for their forecasting model by assuring that their model agrees with the solar cycle data for eight solar cycles in reality this is not difficult to achieve with enough model parameters at one s disposal . figure 5 shows the results of integrating the predictor models for two different choices of incorrect parameter . note that even though the predictor has been designed to reproduce the target over a number of cycles and that the predictor is very closely related to the target , there is still a good chance that it can get the next cycle incorrect , with significant errors in ( particularly ) the cycle amplitude . there are also clear variations in the cycle period , which obviously implies that the exact time between successive cycle maxima is also an unpredictable feature of the system . we stress again that any mean - field model of solar activity includes transport coefficients that are still uncertain possibly to an order of magnitude ( and certainly not to @xmath62 accuracy ) . although the incredible success of global and local helioseismology is placing restrictions on the form of the differential rotation and the meridional flows , it is unlikely in the foreseeable future that significant constraints will be put on the transport coefficients or their nonlinear response to the mean magnetic field . having established that there are difficulties in obtaining reliable predictions by fitting mean - field models ( even if the modulation is deterministic in origin ) , it is of interest to determine whether or not more reliable predictions could be obtained by utilising more general timeseries analysis techniques . in order to reconstruct an attractor from a given timeseries , it is necessary to define a corresponding phase space . there are various ways of doing this , but given ( any ) discrete timeseries , @xmath63 , in which the data is sampled at intervals of @xmath64 , the vector @xmath65 \label{eqn:10}\ ] ] defines a point in a @xmath66-dimensional `` embedded '' phase space ( see , e.g. , farmer & sidorowich 1987 ; casdagli 1989 ) . given a time @xmath67 , the idea of a prediction algorithm is to find a mapping @xmath68 such that @xmath69 gives a good approximation to @xmath70 . the predictive mapping technique that is used here uses a local approximation method ( see , e.g. , casdagli 1989 ) , which considers the behaviour of the nearest neighbours , in phase space , to @xmath71 . by using a least squares fit , the subsequent evolution of each of these neighbouring points in phase space is used to construct a piecewise - linear approximation to the predictive map , @xmath68 . this approximate mapping can then be applied to @xmath71 to obtain an estimate for @xmath72 . this algorithm can then be repeated to find estimates for @xmath73 and subsequent points . the optimal value for @xmath66 can be determined by minimising the error of this predictive algorithm over the known segment of the timeseries . the results of applying this predictor algorithm to the target solution are also shown in figure 5 , where the timeseries predictions are shown as crosses . the prediction is started from the cycle maximum before the mean - field predictor diverges from the target . longer training timeseries lead to a more densely - populated reconstructed attractor , which increases the probability of making more accurate predictions . however , rather than using the entire target run , these predictions are based upon ( approximately ) @xmath74 cycles this will give a fairer comparison between these results and timeseries predictions that are based upon the real sunspot data . the application of the algorithm to earlier segments of the timeseries suggests that a value of @xmath75 is required in order to minimise predictive errors . as can be seen from figure 5 , this algorithm appears to predict the magnitude of the maximum of the following cycle to a reasonable degree of accuracy , although the predictions subsequently diverge from the target . whilst neither of these techniques are capable of producing reliable long - term predictions , these results do suggest that for the short - term prediction of solar magnetic activity , timeseries analysis techniques may provide a viable alternative to predictions based simply upon mean - field dynamo models ( provided stochastic effects can be neglected ) . solar magnetic activity arises as a result of a hydromagnetic dynamo that much we believe to be true . as yet , there is no consensus on the location of the dynamo , the dominant nonlinear or stochastic effects , or even the fundamental processes that are responsible for the operation of such a dynamo . although plausible mechanisms have been proposed , as yet none of these are entirely satisfactory . against this background , there is a drive to be able to predict solar activity with greater accuracy , due to the importance of this activity in driving solar events . what we have demonstrated here is that no meaningful predictions can be made from illustrative mean - field models , no matter how they are constructed . if the mean - field model is constructed to be a driven linear oscillator then the small stochastic effects that lead to the modulation will have an extremely large effect on the basic cycle and make even short - term prediction extremely difficult . the second scenario , where the modulation arises as a result of nonlinear processes rather than stochastic fluctuations , is clearly a better one for prediction though here too , prediction is fraught with difficulties . owing to the inherent nonlinearity of the dynamo system , long - term predictions are impossible ( even if the form of the model is completely correctly determined ) . furthermore , even short - term prediction from mean - field models is meaningless because of fundamental uncertainties in the form and amplitude of the transport coefficients and nonlinear response . any deterministic nonlinear model that produces chaotically modulated activity cycles will be faced with the same difficulties . the equations that describe dynamo action in the solar interior are known to be nonlinear partial differential equations the momentum equation is nonlinear in both the velocity and the magnetic field . one indication of the role played by nonlinear effects in the solar dynamo is the presence of cyclic variations in the solar differential rotation ( the `` torsional oscillations '' ) . furthermore estimates of the field strength at the base of the convection zone consistent with the observed formation of active regions yield fields of sufficient strength ( @xmath76 g ) for the nonlinear lorentz force to be extremely significant , whilst the flows are vigorously nonlinear and turbulent . it therefore seems extremely unlikely that the dynamics of the solar interior can be described by a forced linear system without throwing away much ( if not all ) of the important physics . in this case it must be argued _ not only _ that this discarded physics is irrelevant to the dynamo process but also that the parameterisation of the unresolved physics should not include a stochastic component , as this would have an extremely large effect on such a relinearised system . it is certainly tempting to try to use the observed magnetic flux at the solar surface as an input to a model for prediction ( whether nonlinear or stochastic , mean - field or full mhd ) . certainly any fully consistent solar activity model constructed in the future should be capable of reproducing the observed pattern of magnetic activity at the solar surface , although this will require a complete understanding not only of the generation process via dynamo action , but also the processes which lead to the formation and subsequent rise of concentrated magnetic structures from the solar interior to the surface . however it is not clear what role the flux at the solar surface plays in the basic dynamo process . is it inherent to the process ( as modelled by flux transport dynamos ) or simply a by - product of the dynamo process that is occurring deep within the sun ? estimates suggest that between @xmath77 and @xmath78 of the solar flux generated in the deep interior makes it to the solar surface ( e.g. galloway & weiss 1981 ) . for the flux at the solar surface to be the key for dynamo action , it must be explained why the majority of the magnetic flux that resides in the solar interior plays such a little part in the dynamics ( to such an extent that it does not even appear as a small stochastic perturbation to the large - scale flux transport dynamo ) . finally it is important to stress that _ even if _ a model has been tuned so as to reproduce results over a number of solar activity cycles , then there is a good chance of error in the prediction for the next cycle . any advection - diffusion system in which one is free to specify not only the sources and the sinks but also the transport processes can be tuned to reproduce any required features of activity . moreover , the formulation of a prediction in terms of a parameterised mean - field model does not inherently put the prediction on a sounder scientific basis than a prediction based on methods of timeseries analysis alone ( some of which use very sophisticated mathematical techniques ) . this , of course , is not to say that any given prediction from such a model will be incorrect , just that the basis for making the prediction has no strong scientific support .	we discuss the difficulties of predicting the solar cycle using mean - field models . here we argue that these difficulties arise owing to the significant modulation of the solar activity cycle , and that this modulation arises owing to either stochastic or deterministic processes . we analyse the implications for predictability in both of these situations by considering two separate solar dynamo models . the first model represents a stochastically - perturbed flux transport dynamo . here even very weak stochastic perturbations can give rise to significant modulation in the activity cycle . this modulation leads to a loss of predictability . in the second model , we neglect stochastic effects and assume that generation of magnetic field in the sun can be described by a fully deterministic nonlinear mean - field model this is a best case scenario for prediction . we designate the output from this deterministic model ( with parameters chosen to produce chaotically modulated cycles ) as a target timeseries that subsequent deterministic mean - field models are required to predict . long - term prediction is impossible even if a model that is correct in all details is utilised in the prediction . furthermore , we show that even short - term prediction is impossible if there is a small discrepancy in the input parameters from the fiducial model . this is the case even if the predicting model has been tuned to reproduce the output of previous cycles . given the inherent uncertainties in determining the transport coefficients and nonlinear responses for mean - field models , we argue that this makes predicting the solar cycle using the output from such models impossible .
the blazar is one of most intensively studied bl lac objects . the redshift determinations ( @xmath4 ) are based on measurements of the size of a marginally detected host galaxy and a location near three galaxies with redshifts of 0.26 . optical variability has been studied by many teams ; the blazar exhibits persistent activity on both long and short ( intraday ) timescales , as shown by , e.g. , @xcite ; , @xcite , and @xcite , and by the results of several webt campaigns . violent polarimetric variability of s5 0716 + 71 , among other blazars , was studied by @xcite . @xcite , using a bayesian approach , suggest that the behavior of polarization in s5 0716 + 714 implies the presence of a number of polarization components having quite short timescales of variations . @xcite reported that a huge optical outburst of s 50716 + 71 in 2008 april was accompanied by a @xmath5 rotation of the position angle of the electric vector ( evpa ) of linear polarization . in this paper , we report the results of our observations of s5 0716 + 71 during a major optical outburst in 2011 october , which coincided with an unprecedented @xmath1-ray outburst @xcite . we present our observations in [ obs ] , describe the phenomenological model in [ sec : model ] , and apply it to the data and draw conclusions in [ discuss ] . the observations reported here were collected as a part of a long - term multi - wavelength study of blazars that follows nonthermal activity at optical and @xmath1-ray frequencies . an overview of this program is given in @xcite . we obtained observational data at the 70-cm azt-8 reflector of the crimean astrophysical observatory , 40-cm lx-200 telescope in st . petersburg , both equipped with identical photometers - polarimeters based on st-7 ccds , and 1.8-m perkins telescope of lowell observatory using the prism camera . the photometric measurements in @xmath0 band were supplemented by observations at the 2-m liverpool telescope at la palma , canary islands , spain . polarimetry at azt-8 and the perkins telescope was made in cousins @xmath0 band , and in white light at lx-200 , with effective wavelength close to @xmath0 . the photometric errors do not exceed @xmath6 . instrumental polarization was found via stars located near the object under the assumption that their radiation is unpolarized . the galactic latitude of s5 0716 + 714 is @xmath7 and @xmath8 , so that interstellar polarization ( isp ) in this direction is less than 0.5% . to correct for the isp , the mean relative stokes parameters of nearby stars were subtracted from the relative stokes parameters of the object . this accounts for the instrumental polarization as well . the errors in degree of polarization are less than 1% ( in most cases less than 0.5% ) , while the evpa is determined with a precision of 12 degrees . our regular photometric and polarimetric monitoring of s5 0716 + 71 started in 2005 . since then , we have obtained more than 1000 data points almost uniformly spread over a 7-year interval , although we intensified our observations during the periods of highest activity of the source . figure [ s5total ] presents the flux and polarization behavior of for 20052011 . we supplement this plot with a panel showing the @xmath1-ray light curve from _ fermi _ lat open access data in order to show that the most prominent @xmath1-ray activity ever recorded for this source was observed during the october 2011 optical outburst . in fig . [ s5total2011 ] we show a blowup of the last six months of 2011 . table 1 lists the photometric and polarimetric data of s5 0716 + 71 around the time of the outburst . -ray flux density ( purple points mark upper limits ) , @xmath0-band flux , and degree and position angle of optical polarization of s5 0716 + 71 during 20052011 . red horizontal line in the bottom panel corresponds to the position angle of the radio jet in 2011.,width=283 ] centered around 2011 outburst . red arrows on @xmath0 flux panel mark positions of quasi - periodic optical flares.,width=283 ] cccrcccc 55823.516 & 13.496 & 0.012 & 20.11 & 0.41 & 246.7 & 0.5 & lx-200 + 55826.543 & 13.002 & 0.003 & 8.64 & 0.17 & 262.9 & 0.5 & azt-8+st7 + 55827.554 & 13.012 & 0.002 & 16.90 & 0.24 & 258.7 & 0.4 & azt-8+st7 + 55828.547 & 13.582 & 0.003 & 5.51 & 0.40 & 257.6 & 2.1 & azt-8+st7 + 55829.544 & 13.542 & 0.004 & 16.75 & 0.28 & 235.6 & 0.4 & azt-8+st7 + 55830.560 & 13.598 & 0.003 & 9.77 & 0.50 & 243.0 & 1.4 & azt-8+st7 + 55831.574 & 13.219 & 0.002 & 3.17 & 0.16 & 254.0 & 1.5 & azt-8+st7 + 55832.531 & 13.373 & 0.003 & 11.66 & 0.41 & 269.8 & 1.0 & lx-200 + 55833.533 & 13.361 & 0.003 & 20.14 & 0.32 & 266.9 & 0.4 & azt-8+st7 + 55836.488 & 13.384 & 0.003 & 19.93 & 0.87 & 258.1 & 1.2 & lx-200 + the optical light curve of s5 0716 + 71 displays violent variability on both long ( months - years ) and short ( days - weeks ) timescales . changes in the degree of polarization seem to occur erratically : within the time interval 20052009 it varied from 0 to 15% , while later the mean level of polarization increased , reaching a record value of 36% on the night of 2011 may 14 ( jd 2455696 ) . the evpa also exhibited prominent changes . to solve the @xmath9 ambiguity , we have added / subtracted @xmath10 each time that the subsequent value of evpa is @xmath11 less / more then the preceding one . the bl lac object s5 0716 + 71 is monitored by the boston university group with the very long baseline array ( vlba ) at 43 ghz within a sample of bright @xmath1-ray blazars with an angular resolution of @xmath120.1 mas , which corresponds to 0.45 pc at the source distance ( @xmath13 , @xmath14 , @xmath15 ) . the vlba data were calibrated and imaged in the same manner as discussed in @xcite . we fit the vlba data at 16 epochs from 2010 december to 2012 may by a model consisting of a number of components with circular gaussian brightness distributions . during this period we identify 6 components , @xmath16 , and @xmath17 , in addition to the core , @xmath18 ( fig . [ comp_int ] ) , [ 0716_pfmaps ] ) , which is , presumably , a stationary feature located at the southern edge of the jet in the images . figure [ 0716_components ] shows the evolution of the distance of knots from the core , while figure [ 0716_traj ] plots the trajectories of the knots . . new superluminal knot k3 appeared on mjd 55850 , which coincides within 1-@xmath19 uncertainty with the time of the highest peak in the @xmath1-ray light curve.,width=283 ] knots @xmath20 and @xmath17 move to the northeast along slightly different trajectories with apparent speeds of [email protected] c and 21@xmath212 c , respectively . an apparent speed of @xmath1220 c seems to be common in the parsec scale jet of s5 0716 + 71 . knot @xmath20 undergoes both an acceleration along the jet @[email protected] mas yr@xmath23 and perpendicular to the jet @[email protected] mas yr@xmath23 . an extrapolation of the motion of @xmath17 back to the core suggests that it passed through the core on mjd 55850@xmath2110 . figure [ 0716_pfmaps ] shows total and polarized intensity images of the quasar while @xmath17 emerged from the core . features @xmath25 , @xmath26 , @xmath27 , and @xmath28 are rather stable in distance with respect to the core , although they show scatter in position angle that exceeds the uncertainties of the modeling . @xcite found that `` stationary hot spots '' are a common characteristic of compact jets , with the majority of such features located within a range of projected distances of 1 - 3 pc from the core . these authors propose three categories of models for stationary components in supersonic jets : a ) standing recollimation shocks caused by imbalances between the pressure internal and external to the jet ; b ) sites of maximum doppler beaming where the jet points most closely to the line of sight ; and c ) stationary oblique shocks , where the jet bends abruptly . we conclude that knots @xmath25 , @xmath26 , and @xmath27 fall most likely in category @xmath29 , since they are quasi - stationary with an observed `` lifetime '' of several months to a year . in addition , according to numerical simulations by @xcite , when a moving knot passes through a standing recollimation shock , the components blend into a single feature and split up after the collision with no changes in proper motion of the moving knot . this scenario seems to be observed in the case of @xmath20 approaching @xmath26 ( fig . [ 0716_components ] ) . knot @xmath27 , however , could be a trailing component that forms behind a strong shock @xcite , since it is observed just after the appearance of the fast moving knot @xmath17 . the latter might change the internal pressure in the jet , resulting in the disappearance of knots @xmath25 and @xmath26 ( see fig . [ 0716_components ] ) . knot @xmath28 may belong to category @xmath30 , since the position angle of @xmath28 is different from that of @xmath25 , @xmath26 , and @xmath27 ( fig . [ 0716_traj ] ) and the kinematics of moving knot @xmath20 changes in the vicinity of @xmath28 . in the case of a stationary feature produced at a bend in the jet , the proper motion of a moving knot approaching the stationary feature should decrease before and increase after the knot passes through the bend @xcite . such a behavior can be inferred from figure [ 0716_components ] when knot @xmath20 approaches and recedes from knot @xmath28 . the @xmath1-ray emission from s5 0716 + 71 was detected throughout the entire period of _ fermi _ observations starting in 2008 august , although most of the time at a level @xmath31 . we analyze the _ fermi _ lat data for the period of 2011 january - november with 2-day binning , as well as with 1-day binning and a 6-hour shift , during the huge @xmath1-ray outburst of 2011 october , when the flux from the source was close to @xmath32 . this prevents missing any possible short - lived events , to make the correlation analysis more robust , and to avoid the dependence of the results on the start of the time bins . we use the standard _ analysis software package science tools ` v9r27p1 ` , with the instrument response function ` p7source_v6 ` , the galactic diffuse emission model ` gal_2yearp7v6_v0 ` , and the isotropic background model ` iso_p7v6source ` . the background models include all sources from the 2fgl within @xmath33 . the spectrum of s5 0716 + 71 as well as spectra of all background sources are modelled as a power - law with photon index fixed to the catalog values . we calculate the discrete correlation function ( dcf ) @xcite of the optical and @xmath1-ray flux variations during 2011 . the results are given in fig . optical variations lag those at @xmath1-ray energies by @xmath34 days ( dcf centroid position ) . -ray flux variations of s5 0716 + 714 . positive lag means that optical variations follow those at @xmath1-ray energies.,width=283 ] we suppose , following @xcite , that most of optical photometric and polarimetric variability arises when a compact emission region ( e.g. , a shock wave ) propagates downstream from the black hole , following a spiral path . ( alternatively , the jet could have a helical geometry . ) we notice that prominent optical outbursts are quite often both preceded and followed by minor flares . we interpret this phenomenon as a manifestation of the oscillating doppler beaming of the emission ( lighthouse effect , see ) . the observed series of outbursts correspond to the time intervals when the viewing angle of the shock wave is at a minimum . the main parameters that determine the visible behavior of the outburst are : * jet viewing angle @xmath35 * bulk lorentz factor @xmath36 of the shocked plasma * temporal evolution of the outburst @xmath37 ; the factor @xmath38 is responsible for different time scales of the rise and decline of the outburst ; primed quantities refer to the plasma frame * doppler time contraction in the observer s frame , @xmath39 * shocked plasma compression @xmath40 = ratio of post - shock to pre - shock density * spectral index of the emitting plasma @xmath41 * pitch angle @xmath42 of the spiral motion and helical field * period of the shock s spiral revolution in the observer s frame @xmath43 . the viewing angle of the compact emission region @xmath44 is obtained from the relation : @xmath45 relativistic aberration leads to a change in direction of the normal to the shock front @xmath46 : @xmath47 . \label{eq2}\ ] ] if the _ unshocked _ magnetic field is chaotic upstream of the shock wave , the fractional polarization of the shocked plasma radiation ( see , e.g. , * ? ? ? * ) is : @xmath48 and the position angle is determined by the direction of the projected minor axis of the shock wave : @xmath49 . \label{eq4}\ ] ] the doppler factor @xmath50 is given by @xmath51 the periods of rotation in the source frame @xmath52 and observer s frame @xmath43 are connected through the relation @xmath53 where @xmath54 is the source s redshift . knowledge of @xmath43 ( from the timescale of recurrence of outbursts ) , estimates of @xmath55 and @xmath35 , obtained from analysis of radio images , and evaluation of the spiral s radius @xmath56 allow us to constrain the values of @xmath52 and @xmath42 and to calculate the doppler - contracted time in the observer s frame , @xmath57 . doppler boosting of the radiation from the shocked plasma , as a discrete source , leads to @xmath58 we notice that in most cases when the source is in a quiescent state and no outburst is detected , there remains polarization at the level of @xmath59 . we assume that this is the polarized radiation of the ambient jet , presumably changing on substantially longer timescales ( except possibly for random fluctuations due to turbulence in the jet ) . in the local frame the ratio @xmath60 is of the order of a few percent . the radiation from the undisturbed jet is also subject to doppler boosting , and we ( rather arbitrarily ) set @xmath61 . since the ambient jet is a roughly steady and continuous source , its radiation is amplified by @xmath62 . notice that , in spite of the low value of @xmath63 , in the observer s frame the boosted radiation of the shock may far exceed that of the ambient jet . we add emission from the quasi - stationary features probably standing shocks whose fluxes become significant after the moving shock crosses them . we postulate that the temporal behavior of this added radiation is of the same form as that of the shock : @xmath64 here @xmath65 accounts for the ratio of the ambient jet to shock luminosity , with time delay @xmath66 . in our model we suppose that this radiation is added to the `` constant '' radiation of the jet , but not to its polarized part . finally , the observed radiation of the jet is @xmath67 and total observed flux @xmath68 . we consider a single variable source plus an initially constant source whose polarization vector is displaced relative to the jet direction by an angle @xmath69 ( see , e.g. , * ? ? ? accounting for the contribution of this `` constant '' polarized source ( @xmath70 and @xmath71 ) , we get the absolute stokes parameters : @xmath72 finally , the normalized stokes parameters are given by @xmath73 figure [ model1 ] shows the dramatically different behavior of flux and polarization with all of the above mentioned parameters fixed , allowing only the lorentz factor of the shock to change . we see that the model of a shock propagating in the helical path may in a natural way explain the successions of outbursts , often observed in blazar light curves , as a manifestation of a single event . a variety of observed patterns of photometric and polarimetric behavior may be explained by the interplay of the shock lorentz factor , jet viewing angle , helical path pitch angle , and shape and length of the outburst . we do not take into account the light travel delay caused by the finite size of the shock wave ; if included , this would result in some smoothing of the model light curve and polarization behavior . quasi - periodic flares in the light curves of blazars are expected in models of magnetically launched jets , where the plasma moves with relativistic velocities following helical paths @xcite . in light of this , it is rather surprising that numerous extensive observational campaigns have uncovered only a few cases that could be interpreted as a manifestation of a lighthouse effect from spiral relativistic motion , e.g. , ( @xmath74 , optical @xmath75 band * ? ? ? * ) ; ( @xmath76 , optical @xmath77 bands * ? ? ? ( @xmath78 , optical photometry and polarimetry * ? ? ? * ) ; and ao 0235 + 16 ( @xmath79 in x - rays * ? ? ? radio variability tends to occur on timescales longer than these periods . however , radio mapping provides some picturesque examples of periodic structures that correspond to longer timescales in agn jets , as in the case of ( @xmath80 years * ? ? ? as noted in , the jet might exhibit a fractal helical structure , where geometric scales that increase with distance from the central engine correspond to increasing variability timescales . recent polarimetric studies of the nucleus of @xcite have revealed strong variations of the polarization vector , synchronous with the flux variations , that can be interpreted as a manifestation of a shock propagating through a helical jet . we have obtained densely sampled optical photometric and polarimetric data during an unprecedented @xmath1-ray outburst in s5 0716 + 71 , and analyzed the changes in the structure of the blazar at 43 ghz . we find a @xmath10 rotation of the position angle of linear polarization , which occurred at the epoch of maximum optical brightness . this event would have been missed if our observations did not occur consecutively from 3 sites ( st . petersburg , crimea , arizona ) . this implies that such events may be rather common during blazar flares . figure [ model ] shows the time evolution of the observational parameters ( @xmath1-ray and optical flux , degree of polarization and positional angle ) within the model discussed above . values of the adjustable parameters are listed in table [ tab_1 ] . several of the trends of the data resemble similar features in the model flux and polarization curves . particularly , we see a succession of quasi - periodic flares before and after the main outburst ( marked with vertical arrows in fig [ s5total2011 ] ) and fast oscillation of the degree of polarization around the main flare . this conforms with the model presented here . turbulence may play a role in modulating the behavior ( e.g. , the discrepant flux and degree of polarization near mjd 55675 ) beyond the capabilities of the simplified model proposed here . the @xmath1-ray outburst started @xmath814 before the optical , which corresponds to @xmath82 pc distance in the source frame . the emergence of knot k3 at the same time as the optical/@xmath1-ray outburst leads to the conclusion that these events were co - spatial . however , we note that our model uses a lower lorentz factor of the shock wave ( @xmath83 ) and a wider viewing angle ( @xmath84 ) than that obtained from the radio images ( @xmath85 ) . this could reflect acceleration of the plasma flow in the jet downstream from the regions where the optical/@xmath1-ray outburst took place ( see , e.g. , * ? ? ? we acknowledge the referee s useful comments and suggestions . this work was partly supported by russian rfbr grants 12 - 02 - 00452 , 12 - 02 - 31193 and by nasa fermi guest investigator grants nnx-08av65 g , -11ao37 g and -11aq03 g . the vlba is operated by the national radio astronomy observatory . the national radio astronomy observatory is a facility of the national science foundation operated under cooperative agreement by associated universities , inc . the prism camera at lowell observatory was developed by k. janes et al . at bu and lowell observatory , with funding from the nsf , bu , and lowell observatory the liverpool telescope is operated on the island of la palma by liverpool john moores university in the spanish observatorio del roque de los muchachos of the instituto de astrofisica de canarias , with financial support from the uk science and technology facilities council .	we present the results of optical ( @xmath0 band ) photometric and polarimetric monitoring and very long baseline array ( vlba ) imaging of the blazar along with _ fermi _ @xmath1-ray data during a multi - waveband outburst in 2011 october . we analyze total and polarized intensity images of the blazar obtained with the vlba at 43 ghz during and after the outburst . monotonic rotation of the linear polarization vector at a rate of @xmath2 per night coincided with a sharp maximum in @xmath1-ray and optical flux . at the same time , within the uncertainties , a new superluminal knot appeared , with an apparent speed of @xmath3 . the general multi - frequency behavior of the outburst can be explained within the framework of a shock wave propagating along a helical path in the blazar s jet .
non - thermal x - ray emission has been detected from a steadily growing number of supernova remnants ( snrs ) located in the galaxy . thanks to significant advances in the angular resolution capabilities of modern x - ray observatories , important progress has been made in localizing the origin of this particular type of emission . such localization has revealed that this emission is not associated with any neutron stars observed in projection toward these snrs ( which may or may not be physically associated with the snrs themselves ) : also , this localization is essential for identifying the radiation mechanism responsible for this emission . in the case of six snrs namely sn 1006 @xcite , [email protected] @xcite , [email protected] @xcite , [email protected] @xcite , [email protected] @xcite and [email protected] @xcite the detected x - ray emission is predominately non - thermal . for a subset of these snrs ( sn 1006 , [email protected] , [email protected] , [email protected] and [email protected] ) the non - thermal x - ray emission is localized to narrow filaments of emission located on the leading edge of the snr . other galactic snrs such as cas a , kepler , tycho and rcw 86 feature significant amounts of non - thermal x - ray emission in addition to the typical thermal x - ray emission detected from snrs @xcite . in the cases of these particular snrs , non - thermal x - ray emitting filaments are observed surrounding regions of thermal - dominated x - ray emission . in fact , in the case of cas a ( and likely also to be the case for other snrs as well ) , filaments of non - thermal emission are also detected toward the central regions of the snr : such filaments may be physically located on the edge of the snr but seen toward the center due to projection effects or they may actually be located in the interior of the snr . we note that cas a also exhibits evidence for an unresolved component to its non - thermal x - ray emission : evidence for the existence of this component includes measurements of the total flux from the sum of these filaments ( including both filaments located along the edge of cas a as well as filaments seen toward the interior ) can only account for one - third of the non - thermal x - ray emission detected from the snr by other x - ray observatories . complementary _ xmm - newton _ images of cas a at higher x - ray energies ( corresponding to approximately 10 kev ) suggest that this unresolved component may be associated with the reverse shock or contact discontinuity ( in other words , a region located inside the snr ) . in fact , in an analysis of _ suzaku _ observations of cas a , @xcite found that the peak of hard x - ray emission ( for energies between 11 and 14 kev ) possibly coincides with the location of an observed peak of tev emission from the snr : those authors suggest that this finding indicates that both high - energy hadrons as well as leptons can be accelerated in the reverse shock in an snr , since the tev peak should have a hadronic origin ( see , for example , @xcite ) . in the cases of those snrs which feature x - ray spectra dominated by non - thermal x - ray emission , a diffuse component of x - ray emission of this type is usually not detected . based on data from pointed observations made with the advanced satellite for cosmology and astrophysics ( @xmath4 ) as well as data from the @xmath4 galactic plane survey @xcite , @xcite identified three new galactic candidate x - ray snrs ( [email protected] , [email protected] and [email protected] ) which may also belong to this class and @xcite identified yet another possible member , [email protected] . @xcite also estimated that approximately 20 galactic snrs with x - ray spectra dominated by non - thermal emission may be luminous enough to have been detected by the @xmath4 galactic plane survey . therefore , it is reasonable to expect that even more snrs of this type will be detected and identified as the sensitivities and imaging capabilities of x - ray observatories continue to improve and more x - ray observations of the galactic plane are conducted . synchrotron radiation is a commonly accepted mechanism for the origin of the observed non - thermal x - ray emission detected from snrs . in the case of a synchrotron origin , the observed radiation is being produced by highly - relativisitic cosmic - ray electrons accelerated along the expanding shock fronts of these snrs @xcite and localized to the observed filaments . other processes besides synchrotron radiation namely inverse compton scattering of cosmic microwave background photons and non - thermal bremsstrahlung have also been proposed to account for the observed non - thermal x - ray emission from snrs , but these two other proposed processes both have difficulties in adequately modeling this emission @xcite . a non - thermal bremsstrahlung origin for the emission has been ruled out because the same electron population responsible for such emission should also excite emission lines in the x - ray spectra which are not observed . in the case of a possible inverse compton scattering origin , the predicted slope of the x - ray spectrum produced by this process is much flatter than the slopes of the detected x - ray spectra . additional evidence for a synchrotron origin of this observed x - ray emission stems from qualitatively consistent simple modeling of synchrotron emission from snrs over broad ranges of photon energies , namely the radio through the x - ray @xcite . debate still persists in regards to the diffuse non - thermal x - ray emission detected from cas a : some authors have argued that this emission has a synchrotron origin @xcite while others suggest a non - thermal bremsstrahlung origin @xcite . we note that morphological comparisons between radio and x - ray images of snrs have also been made as part of studies of x - ray synchrotron emission from these sources . these comparisons assume that the radio emission from an snr has a synchrotron origin and have concentrated on higher x - ray energies to avoid confusion with thermal emission from the snr at lower x - ray energies . the results of these morphological comparisons have been mixed . a robust correspondence between x - ray and radio emission along the northeastern rim of sn 1006 has been noted @xcite as well as a broader overlap between x - ray and radio emission from the northwestern rim of [email protected] @xcite : these results support a synchrotron origin for the non - thermal x - ray emission detected from these snrs . however , in the case of [email protected] an anti - correlation between x - ray and radio emission is observed : the northern rim is a strong radio source but weak in x - ray while the eastern and western rims are strong in x - ray but weak in radio @xcite . finally , x - ray and radio emission from the snrs [email protected] . [email protected] and [email protected] do not appear to be strongly correlated @xcite , but such comparisons are complicated by the fact that each of these snrs are weak radio emitters . in table [ summaryobssnrstable ] , we present a summary of observations that have been made with the _ chandra _ x - ray observatory @xcite of the six galactic snrs mentioned previously with x - ray spectra dominated by non - thermal emission . these studies have exploited the high angular resolution ( @xmath2 1@xmath5 ) capabilities of _ chandra _ to image the leading filaments located in the x - ray luminous rims of these snrs . a synchrotron origin for the observed non - thermal x - ray emission from the rims of these snrs is of considerable interest to the cosmic - ray astrophysics community because such emission is consistent with models of synchrotron radiation from cosmic - ray electrons accelerated along the forward shock of snrs @xcite . by thoroughly analyzing this observed emission from the rims of snrs , we can address several fundamental issues regarding cosmic - ray acceleration . for example , we can probe the particular distribution of high - energy electrons along acceleration sites and determine if these electrons are gathered in a diffuse plasma , clumped in filaments or have accumulated immediately before the forward shock . we can also estimate the maximum energy of cosmic - ray electrons accelerated along the shock fronts associated with snrs . because cosmic - ray electrons and protons are expected to be accelerated along the shock in a similar manner @xcite , these studies can also help us to improve our understanding of the acceleration of both leptonic and hadronic cosmic - ray particles . there have been several recent claims that tev emission has been detected from three of the snrs mentioned above , namely sn 1006 @xcite , [email protected] @xcite and [email protected] @xcite with the cangaroo ( collaboration of australia and nippon ( japan ) for a gamma ray observatory in the outback ) series of telescopes , specifically cangaroo - i and cangaroo - ii @xcite . we note that similar observations made with another tev observatory , the high energy stereoscopic system , also known as h.e.s.s . @xcite , have also detected tev emission from sn 1006 @xcite , [email protected] @xcite and [email protected] @xcite . currently , there is no consensus about the physical origin of this tev emission : some authors had argued that this emission is from cosmic microwave background photons which have been upscattered to tev energies via inverse compton scattering interactions with ultrarelativistic cosmic - ray electrons @xcite . in contrast , other authors have instead claimed that this emission is due to neutral - pion decay typically at the site of an interaction between an snr and a molecular cloud @xcite . several authors have assumed the former interpretation and included the emission detected at tev energies from these snrs to model processes related to cosmic - ray acceleration by these sources over extremely broad ranges of energies such as radio through tev @xcite . in @xcite , we studied cosmic - ray electron acceleration by one of the snrs with an x - ray spectrum dominated by non - thermal emission ( [email protected] ) with a spatially - resolved broadband ( 0.5 - 30.0 kev ) analysis of data from observations made with _ rosat _ , _ asca _ , and the rossi x - ray timing explorer ( _ rxte _ ) . in this paper , we investigate the phenomenon of cosmic - ray electron acceleration by snrs through an analysis of x - ray and radio observations of the northwestern rim complex of the snr [email protected] made with _ chandra _ and the australia telescope compact array ( atca ) , respectively . @xcite discovered [email protected] as rx j0852.0@xmath04622 , a hard x - ray source approximately 2@xmath6 in diameter seen in projection against the vela snr in _ rosat _ all - sky survey images : this source has also been referred to as vela z and vela jr . in the literature . _ rosat _ observations @xcite and _ asca _ observations @xcite show that the object is shell - like , with a luminous northwestern rim and less luminous northeastern , western and southern rims . in figure [ fig1 ] we present a mosaicked map of g266.2 - 1.2 that was generated from observations made with the position sensitive proportional counter ( pspc ) aboard _ rosat_. the observed shell - like x - ray morphology of [email protected] is similar to the x - ray morphologies of other snrs in its class , such as sn 1006 , [email protected] , [email protected] and [email protected] . an extreme ultraviolet ( euv ) image made of [email protected] at the wavelength of 83 shows a shell - like morphology that broadly matches the observed x - ray morphology @xcite . since its discovery , the age of [email protected] as well as its distance has been the subject of extensive debate in the literature . much of this debate has centered on determining whether [email protected] is located at a distance comparable to the distance to the vela snr ( estimated to be 250@xmath730 pc @xcite ) or if [email protected] lies at a distance beyond the vela snr . distances as low as @xmath2200 pc and ages as low as @xmath2680 - 1000 years have been suggested based on several different observational results . these results include the following : a high temperature ( over 2.5 kev ) derived from fits to the extracted _ rosat _ pspc spectra ( suggestive of a high shock velocity ) coupled with the large angular size of the snr @xcite ; a claimed detection by the imaging compton telescope ( comptel ) aboard the _ compton _ gamma - ray observatory ( cgro ) of @xmath8-ray line emission at 1.157 mev from the titanium-44 decay chain @xcite ; analyses of extracted x - ray spectra of the northwestern rim of [email protected] ( as observed by _ asca _ and _ xmm - newton _ ) which suggested an excess of calcium @xcite and emission lines produced by scandium and titanium @xcite which are expected from the titanium-44 decay chain ; a putative association between a peak in measured nitrate abundances in antarctic ice cores with historical supernovae @xcite and the scenario of a very recent supernova event producing [email protected] @xcite ; finally , age estimates based on the measured widths of the non - thermal x - ray filaments in the northwestern rim of [email protected] as observed by _ chandra _ and the application of models of synchrotron losses for cosmic - ray electrons accelerated along this rim @xcite . however , other studies and accompanying analyses have produced arguments that indicate that [email protected] is significantly more distant ( and thus older ) : for example , fits to the extracted x - ray spectra of the bright rims of [email protected] as observed by _ asca _ and _ xmm - newton _ revealed column densities that are elevated compared to column densities measured toward the vela snr @xcite . in fact , @xcite and @xcite argued that [email protected] is physically associated with a concentration of molecular clouds known as the vela molecular ridge @xcite which lie at a distance of @xmath2800 - 2400 pc @xcite . other studies have cast doubt on the statistical significance of the claimed detections of x - ray and @xmath8-ray emission lines in extracted spectra from the northwestern rim of [email protected] @xcite as well as associations in general between historical supernovae and peaks in nitrate abundances in antarctic ice cores @xcite . recently , @xcite estimated an expansion rate of [email protected] based on two observations made with _ xmm - newton _ of the northwestern rim of [email protected] that were separated by 6.5 years . those authors derived a low expansion rate of [email protected]% for [email protected] : from this rate they estimated an age of 1.7 - 4.3 @xmath9 10@xmath10 years for the snr and a distance of 750 pc . we will adopt this distance to [email protected] for the remainder of this paper . in contrast to the extensive debate about the age of [email protected] and its true distance , there has been general agreement about the type of supernova that produced this snr , namely a type ib / ic / ii event with a massive stellar progenitor . this conclusion is based on estimates of the current snr shell expansion speed @xcite as well as analyses of a central x - ray source in this snr using data from multiple observatories , including _ @xcite , _ asca _ @xcite , _ bepposax _ @xcite and _ chandra _ @xcite . based on these analyses , it is believed that the central x - ray source cxou j085201.4@xmath0461753 is a radio - quiet neutron star which was formed when the massive stellar progenitor of that snr collapsed . @xcite claimed to have detected an h@xmath11 nebula that may be physically associated with this central source . however , @xcite have argued that the 65-millisecond radio pulsar psr j0855@xmath04644 seen in projection toward the southeast edge of [email protected] is instead the neutron star physically associated with this snr . we note that @xcite and @xcite claimed to have found optical emission associated with the outer edge of [email protected] , arguing that the filamentary nebula rcw 37 also known as ngc 2736 and the pencil nebula and the x - ray structure labeled knot d / d " by @xcite are associated with [email protected] . @xcite dispute this result , arguing that both rcw 37 and d / d are instead part of the vela snr . in this paper , we present a detailed spatially - resolved spectroscopic x - ray study of fine structure in the luminous northwestern rim complex of [email protected] as revealed by observations made with _ chandra_. this rim is known to be resolved into both a leading rim and a trailing rim @xcite . these _ chandra _ observations have been previously analyzed by @xcite : in the present paper , we extend this work in two ways . firstly , we have measured the widths of the rim structures at more locations . secondly , we also incorporate radio observations of the northwestern rim complex to help constrain synchrotron models used to fit x - ray spectra extracted for seven regions within the complex . the organization of this paper is as follows : in section [ obssection ] , we describe the observations ( and the corresponding data reduction ) made of the northwestern rim complex of [email protected] with _ ( section [ chandrasubsection ] ) and the atca ( section [ atcasubsection ] ) . the data analysis and the results are presented in section [ dataresultssection ] where we describe the observed fine x - ray emitting structures in the rim complex and the measured widths of the observed structures ( section [ rimstructuresubsection ] ) as well as fits to the extracted spectra of several regions in the rim complex using different models for non - thermal x - ray emission ( section [ nonthermalsubsection ] ) . in section [ maximumsection ] we present an estimate for the maximum energy of the cosmic - ray electrons accelerated along the northwestern rim complex of [email protected] : we compare this estimate to estimates published for the maximum energies of cosmic - ray electrons accelerated by other snrs . lastly , we present our conclusions in section [ conclusionssection ] . the northwestern rim complex of [email protected] was observed by _ chandra _ between 2003 january 5 - 7 in two separate epochs ( obsids 3846 and 4414 ) : details of these observations are provided in table [ table1 ] . x - ray emission from the rim complex was imaged with the advanced ccd imaging spectrometer ( acis ) camera aboard _ @xcite in very faint mode with a focal plane temperature of @xmath0120@xmath6 celsius . the acis is composed of a 2 @xmath9 2 ( acis - i ) and a 1 @xmath9 6 ( acis - s ) array of ccds . six of these ten detectors ( acis - i2 , -s0 , -s1 , -s2 , -s3 and -s4 ) were live during the observations . each 1024 pixel @xmath9 1024 pixel ccd has a field of view of 8.4 @xmath9 8.4 . the angular resolution of the _ chandra _ mirrors and acis varies over the observed portion of the northwestern rim of [email protected] from approximately @xmath12 at the aimpoint to @xmath13 for a region that is located approximately @xmath14 off axis . the maximum on - axis effective area for the mirrors and the acis - s3 ccd is approximately 670 @xmath15 at 1.5 kev : for energies between 0.4 and 7.3 kev , the on - axis effective area is greater than 10% of this value . the fractional energy resolution ( fwhm / e ) between these energies ranges from about 0.4 to 0.03 respectively . the sensitive energy bands and energy resolutions of the other five ccds used in the two observations are typically worse than those of the acis - s3 ccd . in figure 1 we have plotted the positions of the fields of view of each active acis ccd during this observation of the northwestern rim complex over the complete _ rosat _ image of the snr for the energy range of 1.3 through 2.4 kev . the acis data for both observations were reduced using the _ chandra _ interactive analysis of observations ( ciao ) @xcite version 4.0.1 ( caldb version 3.4.3 ) . the steps taken in producing new evt2 files for both observations are outlined as follows : first , the ciao tool ` destreak ` was used to remove anomalous streak events recorded by the acis - s4 chip . next , the ciao tool ` acis_process_events ` was used to generate a new evt2 file ( with the parameter settings of ` check_vf_pha = yes ` and ` trail=0.027 ` implemented as recommended for processing observations made in very faint mode ) . this evt2 file was then filtered based on grade ( where grades=0,2,3,4,6 were selected ) , clean " status column ( that is , all bits set to zero ) and photon energy ( all photons with energies between 0.3 and 10 kev were selected ) . good time intervals supplied by the pipeline were also applied : in addition , a light curve for the observation was prepared and examined for the presence of high background flares during the observation using the ciao tool ` chips ` . no variations were seen at the level of 3@xmath16 or greater : we therefore conclude that background flares do not affect the dataset at a significant level . for the final evt2 files generated for obsids 3846 and 4414 , the effective exposure times weere 39363 seconds and 33917 seconds , respectively , for a total effective exposure time of 73280 seconds . the two evt2 files were combined to produce a single co - added image with an elevated signal - to - noise ratio using the ciao tool ` dmmerge ` . we also extracted spectra for seven regions of interest in the rim complex for both of the observations . specifically , we considered two regions which were sampled by the acis - s2 ccd , four regions which were sampled by the acis - s3 ccd and one region which was sampled by the acis - s4 ccd . for each region , we prepared a pulse - height amplitude ( pha ) file , an ancillary response file ( arf ) and a response matrix file ( rmf ) using the contributed ciao script _ specextract _ : this tool automatically accounts for the build - up of absorbing material on the instruments and adjusts the arfs accordingly . the spectra were grouped to a minimum of 15 counts per bin to improve the signal - to - noise ratio . we performed simultaneous joint fits on the extracted spectra for each region from both images : analyses of the spectra of these regions was conducted using the x - ray spectral analysis software package _ xspec _ version 12.4.0ad @xcite . the results of these spectral analyses are presented in section [ nonthermalsubsection ] . the entire angular extent of [email protected] was observed with the atca on 14 november 1999 in array configuration ew214 at the wavelengths of 20 and 13 cm ( @xmath17=1384 and 2496 mhz , respectively ) . details of these observations are presented and described by @xcite : to summarize , the observations were done in so - called `` mosaic '' mode which consisted of 110 pointings over a 12 hour period . the sources 1934@xmath0638 and 0823@xmath0500 were used for primary and secondary calibration , respectively . the miriad @xcite and karma @xcite software packages were used for reduction and analysis . baselines formed with the sixth atca antenna were excluded , as the other five antennas were arranged in a compact configuration . both images feature a resolution of 120 and an estimated root - mean - square noise of 0.5 mjy / beam . also , both images are heavily influenced by sidelobes originating from the nearby strong radio source ctb 31 ( rcw 38 ) . nevertheless , these images allow us to more efficiently study the larger scale components of this snr . in the present paper , we concentrate on the radio properties of the northwestern rim of [email protected] : the portion of the rim complex that was sampled by the _ observation most closely corresponds to the region denoted as n2 " by @xcite . for a region along this rim with an angular size of 75 square arcminutes , we measure integrated flux densities at the frequencies of 1420 mhz and 2400 mhz of @xmath18@xmath19 = [email protected] jy and @xmath18@xmath20 = [email protected] jy , respectively . complementary observations made of this region at two other frequencies , namely 843 mhz and 4800 mhz @xcite , yield additional flux density estimates of @xmath18@xmath21 = [email protected] jy and @xmath18@xmath22 = [email protected] jy , respectively . based on these flux density values , we estimate a spectral index@xmath23 @xmath24 @xmath17@xmath25 . ] @xmath11 = [email protected] . we also interpolate an integrated flux density at 1 ghz for this rim complex of @xmath18@xmath26 = 0.544 jy and an average surface brightness of @xmath27@xmath28 = 7.25@xmath910@xmath29 jy / arcmin@xmath30 . we will use these estimates for the spectral index and the average surface brightness when fitting the extracted x - ray spectra using synchrotron models as described in section [ nonthermalsubsection ] . in figure [ fig2 ] , we present our reduced , co - added and exposure - corrected _ chandra _ image of the northwestern rim complex of [email protected] for the energy range from 1 to 5 kev . the image has also been smoothed with a gaussian of @xmath31 . we have labeled several salient features of this rim complex , including a thin but prominent leading shock , a leading rim and a trailing rim : these features were noticed by prior studies of this rim complex @xcite and the high angular resolution capabilities of _ chandra _ are essential for a detailed spatially - resolved spectroscopic study of this complex . the x - ray emission from this rim complex is known to be rather hard . in figures [ fig3 ] and [ fig4 ] we present the same image of the x - ray emission with the radio contours overlaid at the frequencies of 2496 mhz and 1384 mhz , respectively . as noted in section [ introductionsection ] , filaments with hard x - ray spectra have been observed along the leading rims of several galactic snrs , including sn 1006 @xcite , cas a @xcite , kepler @xcite , tycho @xcite , rcw86 @xcite , [email protected] @xcite , [email protected] @xcite and [email protected] @xcite . it is generally accepted that these filaments originate from non - thermal ( synchrotron ) emission from cosmic - ray electrons accelerated along the snr shock and that the widths of these filaments can help constrain estimates of both the downstream magnetic field and the maximum energy of the synchrotron - emitting electrons @xcite . in table [ table3 ] we present published estimates of widths for the observed filaments for five of the snrs mentioned above : we have also adopted distance estimates to calculate the corresponding linear widths of the filaments as well . remarkably , the ranges of values for the filament widths are rather consistent with each other ( though it should be noted that estimates of distances to galactic snrs remain significantly uncertain ) : similar findings were reported by @xcite and @xcite . this result suggests that a narrow range of widths of the filaments of snrs ( particularly the young snrs considered here ) is dictated by the dynamics of snr evolution . we note that in contradiction the filament widths seem to be much larger for [email protected] by nearly an order of magnitude for an assumed distance of 6.3 kpc @xcite , but the contradiction disappears for the much closer distance of 1 kpc claimed by other authors @xcite . to compare the widths of the observed rims seen in the northwestern rim complex of [email protected] by our _ chandra _ observation with the widths of the filaments seen in the other snrs , we performed the following fitting procedure . we note that this work extends the analysis conducted by @xcite of filaments associated with the rim complex by analyzing a total of eight regions ( compared with three presented in that work ) and regions located on the acis - s2 and -s3 chips ( instead of just the acis - s3 chip ) . at several locations along the bright northwestern rim ( see figure [ figimg ] ) , we characterized the length scales over which the x - ray surface brightness increases from the nominal level outside the snr to the peak value along the rim . at each location , a @xmath32 ( 100 pixel @xmath9 400 pixel ) rectangular extraction box was defined . this particular box size was chosen to strike a balance between choosing a size small enough to avoid artifically broadening the width because the shock front has some curvature ( that is , it is not entirely linear ) but large enough to contain enough photons for detailed spatial fitting . the rotation angles of the boxes were adjusted until the long axes of the boxes appeared to be perpendicular to the filament : next , the declination coordinates of the centers of the boxes were adjusted until these centers were within one pixel of the locations of the peak surface brightness . once these criteria were met , the iterative fitting process began : the events in each box were extracted and a one - dimensional histogram along the long axis of the box was created ( e.g. see fig . [ fighist ] ) . each histogram was fit with the function @xmath33 where @xmath34 is the total number of events at @xmath35 , @xmath35 is the coordinate along the long direction of the box ( @xmath35 is larger at larger radii ) , @xmath36 is the location of the peak surface brightness , @xmath37 is the length scale over which the surface brightness increases from the nominal level `` outside '' the remnant to the peak value at the rim , @xmath38 is the length scale over which the surface brightness decreases `` inside '' the snr from the peak value at the rim , @xmath39 is the peak number of events and finally @xmath40 is the nominal number of background events . the five free parameters in the fits include @xmath36 , @xmath37 , @xmath38 , @xmath39 and @xmath40 . to ensure that the length scales @xmath37 and @xmath38 are not artificially increased by using boxes whose long sides are not perpendicular to the rim , several rotation angles were used . the rotation angles @xmath41 were varied from @xmath42 to @xmath43 in one degree steps . the angles @xmath44 represent the best initial guesses at the rotation angles of the boxes . at each rotation angle , the data were re - extracted , histogramed and fitted as described above . at each location on the rim , the best - fit rotation angle @xmath45 was obtained by fitting the results of @xmath37 versus @xmath41 with a quadratic function of the form @xmath46 ( e.g. see figure [ figangle ] ) . the angles @xmath45 at which the quadratic functions have minima and the best - fit values for @xmath37 at these angles are listed in table [ table4 ] . the boxes are oriented in figure [ figimg ] at the best - fit angles @xmath45 . as shown in figure [ figoffaxis ] , there is a correlation between the lengths @xmath37 and the off - axis angles @xmath47 , which correspond to the off - axis angles of the center of the extraction box . this correlation demonstrates that the point - spread function of the _ chandra _ mirrors contributes significantly to the fitted length scales at least for the locations that are far from the optical axis . the length scales @xmath37 of regions b and c may be relatively free of the effects of the mirrors . in this case , the actual length scale of the filaments may be about 7 pixels ( about @xmath48 ) . the best - fit values for @xmath38 and @xmath49 , which are obtained in the same way as the values for @xmath37 and @xmath45 , are not listed in table [ table4 ] because the length scales @xmath38 are not necessarily the length scales over which the x - ray surface brightness decreases inside the leading shock . it is very likely that the geometry of the snr may not be as simple as a thin spherical shell , which would make estimates of @xmath38 less physically meaningful . some of the emission that appears to be `` inside '' the snr in figure [ figimg ] may be part of the same irregular surface that produces the leading shock . this emission may appear to be inside the remnant because it is seen in projection . any emission from an irregular surface that is seen in projection can ( and did ) result in artificially large fitted values for the length scales @xmath38 . the length scales @xmath37 should be relatively immune to such effects . in table [ table5 ] we list our derived length scales @xmath37 ( in both arcseconds and parsecs ) for the eight analyzed filaments and compare these with estimates for length scales of the filaments associated with five historical snrs that were measured and presented by @xcite and @xcite . lastly , we note the recent work of @xcite who claim that the widths of these filaments are tied to the rapid decline in the strength of the magnetic field downstream from the turbulence rather than the energy losses of the radiating electrons . as described in section [ obssection ] , we have extracted spectra from seven regions in the rim complex and in table [ table6 ] we list relevant properties of these regions . with the _ a priori _ knowledge that the x - ray emission from the rim complex is predominately non - thermal , we attempted to fit the extracted spectra using three different non - thermal models , described below . to create background spectra for the purposes of our spectral modeling , we extracted spectra from regions located just ahead of the leading edge of the rim complex on each chip . this choice of background helps us to account for diffuse thermal emission seen from the vela snr in projection against the rim complex itself @xcite . in all cases , we have used the photoelectric absorption model phabs @xcite to account for the absorption along the line of sight to the northwestern rim complex . the regions considered for spectral analysis ( both source regions and the background regions ) are depicted in figure [ g266regionsfig ] . we note that this work extends previous spectral analysis of this rim complex that was presented by @xcite by considering the properties of regions sampled by the acis - s2 , -s3 and -s4 chips instead of just the acis - s3 chip . the first model we considered was a simple power - law . this model has traditionally been used to fit the x - ray spectra extracted from _ rosat _ and _ asca _ observations of the x - ray luminous rims of galactic snrs that feature non - thermal emission , such as sn 1006 @xcite and [email protected] @xcite , as well as [email protected] @xcite . the typical values for the photon indices derived from fits with power - law models to the spectra of these bright rims are @xmath50 @xmath2 2.12.6 . in the case of [email protected] , @xcite derived a photon index of [email protected] when fitting the spectrum of the northwestern rim , and photon indices of @[email protected] and @[email protected] for fits to the extracted spectra of the northeastern rim and the western rim , respectively . in their spectral analysis of dataset obtained from the _ xmm - newton _ observation of this rim complex , @xcite derived a photon index of @xmath50 @xmath2 2.6 for fits to the higher energy emission ( @xmath51 @xmath52 0.8 kev ) emission from the entire complex : this power - law component was combined with thermal components ( to model emission separately from the vela snr and from [email protected] ) for modeling the entire broadband emission ( 0.2 @xmath53 @xmath51 @xmath53 10.0 kev ) . the results of our fits to the spectra of the seven regions in the rim complex using this model are given in table [ table7 ] : for each region we give the best fit values for the column density @xmath54@xmath55 , photon index @xmath50 , the normalizations of each fit and for each obsid ( along with 90% confidence ranges for each of these parameters ) , the ratio of the values of the @xmath56@xmath30 to the number of degrees of freedom ( i.e. , the reduced @xmath56@xmath30 value ) and the absorbed and unabsorbed fluxes . we have performed our fits over the energy range of 1.0 through 5.0 kev for each extracted spectra . statistically acceptable fits have been found for all regions using the power - law model . our derived photon indices range from approximately 2.38 through 2.70 ( in broad agreement with the photon index value measured for the whole complex by @xcite ) and differences are seen in the values of the photon index for different regions . several regions in the rim complex may actually be physically associated based on similar values for the derived photon indices : one pair of plausibly associated regions are regions # 3 and # 7 , both of which feature indices of @xmath50 @xmath2 2.38 ( the lowest values of photon indices for any of the regions in the rim complex ) and appear to form a clearly - defined leading shock . similarly , regions # 1 and # 4 appear to form a feature , seen in projection which trails behind the leading shock ( though region # 1 is a leading structure in the portion of the rim complex sampled by the acis - s2 chip ) . finally , regions # 2 and # 6 form a trailing rim that is even more clearly distinct from the leading shock ; we suspect that this rim also appears to lie interior to the leading shock simply because of projection effects . in their analysis of _ xmm - newton _ observations of this snr , @xcite first mentioned the presence of this trailing rim though the superior angular resolution of _ chandra _ more clearly resolves this feature as distinct from the leading shock , as noted by @xcite . we interpret the slight range of values of the photon index to possibly indicate the presence of variations in the shock conditions of accelerated electrons within the complex : the lower index of the leading rim indicates that cosmic - ray electrons are being accelerated to higher energies than at the sites of the regions with higher indices . we will continue to discuss variations in the shock conditions in different portions of the rim complex when we describe our spectral fits with the synchrotron models below . for illustrative purposes , in figures [ shockspectrafig ] and [ confconfig ] we present the extracted spectra of one region ( namely region # 3 ) as fit with the phabs@xmath9power law model and confidence contours for this fit , respectively . the other two models that we used to fit the extracted spectra , srcut and sresc , assume a synchrotron origin for the hard x - ray emission observed from the snr . both of these models have been described in detail elsewhere but we provide brief descriptions of each one here . the srcut model @xcite describes a synchrotron spectrum from an exponentially cut - off power - law distribution of electrons in a uniform magnetic field . this model assumes an electron energy spectrum @xmath54@xmath57(@xmath51 ) of the form @xmath58 where @xmath59 is the normalization constant derived from the observed flux density of the region of the snr at 1 ghz , @xmath50 is defined as 2@xmath11 + 1 ( where @xmath11 is the radio spectral index ) and @xmath51@xmath60 is the maximum energy of the accelerated cosmic - ray electrons . the sresc model @xcite describes a synchrotron spectrum from an electron distribution limited by particle escape above a particular energy . this model describes shock - accelerated electrons in a sedov blast wave which are interacting with a medium with uniform density and a uniform magnetic field . the sresc model takes into account variations in electron acceleration efficiency with shock obliquity as well as post - shock radiative and adiabatic losses . @xcite and @xcite have described successfully applying the sresc model to fit extracted x - ray spectra from the bright rims and the interior of sn 1006 using data from observations made by both _ asca _ and _ rxte_. a crucial feature of both the sresc model and the srcut model is that a resulting fit made with either of these models can be compared with two observable properties of an snr , namely its flux density at 1 ghz and radio spectral index @xmath11 . we note that one of the fit parameters of these two models is the cutoff frequency @xmath17@xmath60 of the synchrotron spectrum of the accelerated cosmic - ray electrons . this frequency is defined as the frequency at which the flux has dropped from a straight power law by a factor of ten for srcut and by a factor of six for sresc . we can express @xmath17@xmath60 as ( see , for example , @xcite ) @xmath61 where @xmath62@xmath63 is the magnetic field strength of the snr in @xmath1 g and it is assumed that the electrons are moving perpendicular to the magnetic field . based on the value for @xmath17@xmath60 returned by the sresc and srcut models as well as its normalization @xmath59 , we can model the synchrotron spectrum of the shock - accelerated cosmic - ray electrons associated with snrs and estimate the maximum energy @xmath51@xmath60 of the shock - accelerated electrons . we emphasize here that to make this estimate of the cutoff energy , either the magnetic field needs to be measured or a specific value for the magnetic field needs to be assumed . in applying the sresc and srcut models , the observed radio properties of the snr namely the flux density at 1 ghz and @xmath11 are needed to constrain the x - ray spectral fits in a meaningful manner . like many of the other characteristics of [email protected] that were described in section [ introductionsection ] , the radio properties of this snr have proven to be controversial as well . to investigate the radio properties of this snr , @xcite used data from the deep continuum survey of the galactic plane made with the parkes 64 meter telescope at a frequency of 2.4 ghz and with an angular resolution of 10.4 arcmin as conducted and described by @xcite . @xcite combined this dataset with additional 1.42 ghz data obtained with the 30 meter telescope of the instituto argentino de radioastronomia at villa elisa , argentina : the half - power beamwidth of this instrument at 1.42 ghz is approximately 34 arcminutes . based on these observations , @xcite claimed to have detected extended limb - brightened features associated with this snr at the frequencies of 2.4 and 1.42 ghz and argued that the radio morphology of this snr closely corresponds to the x - ray morphology . different conclusions were reached by @xcite , however , who performed their own analysis of the 2.4 ghz radio data and also considered data from new radio observations made at 1.40 ghz and 4.85 ghz with the parkes telescope : the latter data was obtained as part of the parkes - mit - nrao ( pmn ) survey @xcite and the angular resolutions of these observations were 14.9 and 5 arcminutes , respectively . @xcite argued that the extended features described by @xcite are more likely to be associated with the vela snr than with [email protected] itself . by considering the radio emission detected within a circle corresponding to the x - ray extent of [email protected] , @xcite measured a spectral index @xmath11 = [email protected] ( where @xmath18@xmath23 @xmath24 @xmath17@xmath25 ) for the northern half of the snr using the method of t - t " plots as described by @xcite and argued that this spectral index value was applicable to all of [email protected] . they also estimated the integrated fluxes from this snr at the frequencies of 2.42 and 1.40 ghz to be 33@xmath76 and 40@xmath710 jy , respectively . by extrapolating their measured integrated flux to a frequency of 1 ghz using the above spectral index value , @xcite estimated the integrated flux from [email protected] at that frequency to be 47@xmath712 jy and from this value calculated an average surface brightness of this snr at 1 ghz to be @xmath64@xmath26 to be [email protected] @xmath9 10@xmath65 w m@xmath66 hz@xmath67 sr@xmath67 = [email protected] @xmath9 10@xmath68 jy sr@xmath67 . the broadband radio observations presented by @xcite of [email protected] more clearly revealed a shell - like radio morphology for this snr : in particular , enhanced radio emission at the northwestern rim complex . using the average surface brightness calculated for this rim complex at 1 ghz ( see section [ atcasubsection ] ) and the measured angular sizes of each region , we calculated the corresponding flux density at 1 ghz for that region . the calculated flux densities for each region are provided in table [ table6 ] . we then performed our spectral analysis with the sresc and srcut models using these calculated flux densities for the normalizations . these frozen normalizations are necessary to help make meaningful interpretations of fits to the x - ray spectra of these features assuming a synchrotron origin to the emission . certainly , local variations are expected in the corresponding flux density at 1 ghz for each region : newer radio observations made with higher angular resolution are required to clearly identify such variations and measure the corresponding flux densities . in tables [ table8 ] and [ table9 ] we present the results of our fits to the extracted spectra using the srcut and the sresc model . for each fit we have listed the fit parameters including column density @xmath54@xmath55 , spectral index @xmath11 , cutoff frequency @xmath17@xmath60 ( along with 90% confidence intervals for each of these parameters ) , the normalization value used , the ratio of @xmath56@xmath30 to the number of degrees of freedom for each fit and the corresponding absorbed and unabsorbed fluxes of each region . based on the the fit values for the cutoff frequencies derived from fitting the srcut model to the spectra of the seven regions , we have used equation [ eeqn ] to calculate the maximum energy @xmath69 of cosmic - ray electrons accelerated at each of these regions . following @xcite and @xcite , we have assumed a field strength @xmath70 g , which is expected to be a lower limit on the actual magnetic field strength . in table [ table10 ] we list our computed estimates for @xmath69 for each region and each assumed normalization : our estimates for @xmath69 for the different regions range from 30 to 40 tev with the highest estimated value corresponding to the portion of the leading shock that corresponds to region # 7 . the higher observed value for @xmath69 for this region may indicate that cosmic - ray acceleration is the most robust at this location and that cosmic - ray electrons are accelerated to their highest energies in the narrow regions along the shock front . once again we point out that the _ chandra _ observations have revealed a possible range of spatial variations in the acceleration of cosmic - ray electrons along the entire rim complex . in tables [ table11 ] and [ table12 ] we present listings of published estimates for the cutoff frequencies of cosmic - ray electrons accelerated by a sample of snrs in the galaxy and the large magellanic cloud ( lmc ) , respectively : in each case these estimates have been derived using the srcut model . most of these estimates have been taken from @xcite and @xcite ( tables [ table11 ] and [ table12 ] , respectively ) . we have included in this list several snrs ( like kes 73 , 3c 396 , [email protected] and rcw 103 ) considered by @xcite which are now known to be associated with central x - ray sources which contribute confusing hard emission that was not spatially resolved from the snr itself by the _ asca _ observations . therefore , these estimates for @xmath69 must be interpreted with caution . in addition , we have augmented the lists from those two papers with published estimates from other papers which have considered other snrs . using equation [ eeqn ] and again assuming a magnetic field @xmath70 g , we have calculated a value for @xmath69 for each snr . lastly , we have added to table [ table11 ] our own estimate for @xmath71 and @xmath72 for [email protected] as described in this paper to help generate an up - to - date listing of values for @xmath69 for snrs both in the galaxy and in the lmc . all of the estimates presented here for values of @xmath69 must be interpreted with a great deal of caution , given known significant uncertainties in our understanding about the true shape of the electron spectrum and about the magnetic field strength . as shown in equation [ synceqn ] , the model srcut is based on the assumption that the electron spectrum is a power law with an exponential cut off . if the cosmic - ray energy density at the forward shock is large enough , then the shock transition region will be broadened @xcite . as a result , cosmic - ray spectra do not have power - law distributions : instead , the spectra flatten with increasing energy @xcite . evidence of such curvature has been reported for cas a @xcite , rcw 86 @xcite and sn 1006 @xcite . if magnetic fields in remnants are @xmath73 g or larger , as has been reported for several young remnants , then the electron spectrum should be steepened due to radiative losses ( e.g. , @xcite ) . furthermore , theoretical analyses presented by other authors @xcite have used an alternative form for the cutoff ( that is , @xmath74 ) where the range of values for @xmath75 have included 1/4 , 1/2 , 1 or 2 ( though @xcite have argued in favor of the use of @xmath76 ) . taken altogether , it is clear that the simple spectral form of the accelerated electrons in equation [ synceqn ] ( as featured in such models as srcut and sresc ) is likely to be a significant over - simplification of the true shapes of the spectra . we comment that the data presently available for [email protected] may not be sensitive enough for very rigorous models of the shape of the electron spectrum . because the synchrotron emissivity function for a monoenergetic collection of electrons is quite broad @xcite , the shape of the synchrotron spectrum will be a very broadly smeared version of the shape of the electron spectrum at the energy where the electron spectrum is cutoff . the synchrotron spectrum can also be broadened if the electron spectrum or magnetic field strength vary within a spectral extraction region . we also comment that the cutoff frequency is only one of the parameters in the fit that control the shape of the spectrum : our fits also include a photon index and an absorption column density and the values of these shape parameters are correlated . if the electron spectrum is curved and if radiative losses are important at the cutoff energy in the electron spectrum , then the synchrotron model is clearly over - simplified and should include additional shape parameters . if the electron spectrum is curved due to the pressure of cosmic rays or synchrotron losses or if @xmath77 , then the estimates for the maximum electron energies in tables [ table11 ] and [ table12 ] are overestimates . furthermore , the assumed magnetic field strength of @xmath78 g is clearly an underestimate for some young snrs , where the magnetic field is known to be closer to @xmath79 g : in such a case , the estimates for the maximum energies given here are certainly overestimates and the true maximum energies may be lower still . despite these reasons to expect that the true electron cut - off energies are lower than the values listed in tables [ table11 ] and [ table12 ] , it is still clear from the augmented tables that no known snr ( either galactic or in the lmc ) appears to be accelerating cosmic - ray electrons to the energy of the knee feature of the cosmic - ray spectrum ( that is , approximately 3000 tev ) , just as pointed out by @xcite in referencing their table 11 . from the work in the present paper , we have established that [email protected] does not appear to be accelerating cosmic - ray electrons to the knee feature either : we note that the values listed for @xmath71 in tables [ table11 ] and [ table12 ] are derived from integrated spectra for these snrs , while we have considered for [email protected] the region with the highest value for @xmath71 . indeed , wide variations in the values for this quantity have been reported for spatially - resolved spectral analyses of snrs like sn 1006 @xcite and cas a @xcite . also , as stated by @xcite , it must be stressed that these maximum values are most likely overestimates given the assumption that most or all of the x - ray flux from an snr is non - thermal in origin . ignoring the contribution of thermal emission is clearly incorrect for snrs which are known to feature emission lines in their x - ray spectra . we note that additional analyses published by other authors have provided _ lower _ limits on the maximum energy of cosmic - ray electrons accelerated by snrs . for example , in analyzing synchrotron emission at infrared wavelengths from cas a , @xcite concluded that the maximum energy of cosmic - ray electrons accelerated by that snr was at least 0.2 tev . given the significant losses through synchrotron radiation which are expected from cosmic - ray electrons as they are accelerated to high energies , it is not very surprising that no snr has been observed to be accelerating cosmic - ray electrons to the energy of the knee feature . the fact that no snr features an @xmath69 value that is near the knee energy is not unexpectedrefle given that the knee feature is more clearly associated with cosmic - ray protons and nucleons . our analyses place no constraints upon the maximum energy of the nuclei . the conclusions of this paper may be summarized as follows : \1 ) we have conducted an analysis of two x - ray observations made with _ chandra _ of the luminous northwestern rim complex of the galactic snr [email protected] : the combined effective integration time of the two observations was 73280 seconds . this snr is a member of the class of galactic snrs which feature x - ray spectra dominated by non - thermal emission , which is commonly interpreted to be synchrotron in origin . we have performed an x - ray spatially - resolved spectral analysis of this rim complex to help probe the phenomenon of non - thermal x - ray emission from snrs and its relationship to cosmic - ray acceleration by these sources . these observations have revealed fine structure in the rim complex , including a sharply defined leading shock and additional filaments . to help constrain the analysis of synchrotron emission from this rim complex , we have also considered radio observations made of [email protected] with atca at the frequencies of 1384 mhz and 2496 mhz . \2 ) we have modeled the length scales of the upstream x - ray emitting structures located along the northwestern rim complex and sampled by the acis - s2 , -s3 and -s4 chips . we compared these length scales with published length scales for other non - thermal x - ray - emitting structures associated with the rims of five historical galactic snrs . the length scales range from 0.02 to 0.08 pc and are comparable to those measured for the older historical snrs . \3 ) we have extracted spectra from seven regions that were sampled using the acis - s array ( namely the acis - s2 , acis - s3 and acis - s4 chips ) and fit the spectra using several non - thermal models , namely a simple power - law and the synchrotron models srcut and sresc . statistically acceptable fits have been derived using each model : however , slight variations are seen in derived fit parameters such as the photon index and cutoff frequency . we interpret the variations of these parameters to possibly indicate differences in how cosmic - ray electrons are accelerated along different sites of the rim complex . we argue that the lower photon index and higher cutoff frequencies derived for the two regions that compose the leading shock ( our regions # 3 and # 7 ) are indicative of cosmic - ray electrons being accelerated to the highest energies at these locations . \4 ) we estimate the maximum energy of cosmic - ray electrons accelerated along this rim complex to be approximately 40 tev . this maximum energy is associated with region # 7 , a region that forms a portion of the leading shock of the rim complex . we have also estimated the maximum energies of cosmic - ray electrons accelerated by other galactic snrs and lmc snrs based on published values of cut - off frequencies . as established in prior works , no known snr appears to be currently accelerating cosmic - ray electrons to the knee energy of the observed cosmic - ray spectrum ( though this knee feature is more applicable for cosmic - ray protons and cosmic - ray nucleons ) . such a result is not unexpected , given that significant synchrotron losses from accelerated cosmic - ray electrons are predicted to limit the maximum energies of these accelerated particles to energies significantly below the energies associated with the knee feature . we thank the referee for many useful comments which helped improve the overall quality of this paper . t.g.p . acknowledges useful discussions with steven reynolds ( regarding the srcut model ) , kelly korreck , jonathan keohane and michael stage . this research has made use of nasa s astrophysics data system : we have also used data obtained from the high energy astrophysics science archive research center ( heasarc ) , which is provided by nasa s goddard space flight center . t.g.p . also acknowledges support for this work from nasa ltsa grant nag5 - 9237 . acero , f. , aharonian , f. a. , akhperjanian , a. g. , anton , g. , barres de almeida , u. , bazer - bachi , a. r. , becherini , y. , behera , b. , beilicke , m. , bernlhr , k. , bochow , a. , boisson , c. , et al . 2010 , , 516 , 62 aharonian , f. a. , akhperjanian , a. g. , aye , k .- m . , bazer - bachi , a. r. , beilicke , m. , benbow , w. , berge , d. , berghaus , p. , bernlhr , k. , bolz , o. , boisson , c. , borgmeier , c. , et al . 2004 , nature , 432 , 75 aharonian , f. a. , akhperjanian , a. g. , bazer - bachi , a. r. , beilicke , m. , benbow , w. , berge , d. , bernlhr , k. , boisson , c. , bolz , o. , borrel , v. , braun , i. , breitling , f. , et al . 2005 , , 437 , l7 aharonian , f. a. , akhperjanian , a. g. , bazer - bachi , a. r. , beilicke , m. , benbow , w. , berge , d. , bernlhr , k. , boisson , c. , bolz , o. , borrel , v. , braun , i. , breitling , f. , et al . 2006 , , 449 , 223 aharonian , f. a. , akhperjanian , a. g. , bazer - bachi , a. r. , beilicke , m. , benbow , w. , berge , d. , bernlhr , k. , boisson , c. , bolz , o , borrel , v. , braun , i. , brown , a. m. , et al . , 2007a , , 661 , 236 aharonian , f. a. , akhperjanian , a. g. , bazer - bachi , a. r. , beilicke , m. , benbow , w. , berge , d. , bernlhr , k , boisson , c. , bolz , o. , borrel , v. , braun , i. , brion , e. , et al . 2007b , , 464 , 235 allen , g. e. , keohane , j. w. , gotthelf , e. v. , petre , r. , jahoda , k. , rothschild , r. e. , lingenfelter , r. e. , heindl , w. a. , marsden , d. , gruber , d. e. , pelling , m. r. & blanco , p. r. 1997 , , 487 , l97 allen , g. e. , petre , r. & gotthelf , e. v. 2001 , , 558 , 739 allen , g. e. , stage , m. d. , & houck , j. c. 2008a , international cosmic ray conference , 2 , 839 allen , g. e. , houck , j. c. & sturner , s. j. 2008b , , 683 , 773 aschenbach , b. , egger , r. & trmper , j. 1995 , nature , 373 , 587 arnaud , k. a. 1996 , astronomical data analysis software and systems v , eds . jacoby , g. and barnes , j. , p17 , asp conf . series volume 101 . aschenbach , b. 1998 , nature , 396 , 141 aschenbach , b. , iyudin , a. f. & schnfelder , v. 1999 , , 350 , 997 aschenbach , b. 2002 , proceedings of the 270th we - heraeus seminar on neutron stars , pulsars and supernova remnants , physikzentrum bad honnef , eds . w. becker , h. lesch & j. trmper , mpe report 278 , 18 ( astro - ph/0208492 ) bamba , a. , ueno , m. , koyama , k. & yamauchi , s. 2001 , , 53 , l21 bamba , a. , yamazaki , r. , ueno , m. & koyama , k. 2003 , , 589 , 827 bamba , a. 2004 , ph.d thesis , kyoto university bamba , a. , yamazaki , r. & hiraga , j. s. 2005a , , 632 , 294 bamba , a. , yamazaki , r. , yoshida , t. , terasawa , t. & koyama , k. 2005b , , 621 , 793 bamba , a. , fukazawa , y. , hiraga , j. s. , hughes , j. p. , katagiri , h. , kokubun , m. , koyama , k. , miyata , e. , mizuno , t. , mori , k. , nakajima , h. , ozaki , m. , et al . 2008 , , 60 , 153 bell , a. r. 1987 , , 225 , 615 benbow , w. , for the hess collaboration , in proceedings of the gamma 2004 symposium on high - energy gamma - ray astronomy , vol . 745 ( aip conference proceedings , eds . aharonian , h. j. v"olk and d. horns ) , 611 berezhko , e. g. & ellison , d. c. 1999 , , 526 , 389 berezhko , e. g. & vlk , h. j. 2008 , , 492 , 695 borkowski , k. j. , rho , j. , reynolds , s. p. & dyer , k. k. 2001 , , 550 , 334 burgess , c. p. & zuber , k. 2000 , astroparticle physics , 14 , 1 cassam - chena , g. , decourchelle , a. , ballet , j. , sauvageot , j .- l . , dubner , g. & giacani , e. 2004 , , 427 , 199 cassam - chena , g. , hughes , j. p. , reynoso , e. m. , badenes , c. & moffett , d. 2008 , , 680 , 1180 cha , a. n. , sembach , k. r. & danks , a. c. 1999 , , 515 , l25 chen , w. & gehrels , n. 1999 , , 514 , l103 combi , j. a. , romero , g. e. & benaglia , p. 1999 , , 519 , l177 duncan , a. r. , stewart , r. t. , haynes , r. f. & jones , k. l. 1995 , , 277 , 36 duncan , a. r. & green , d. a. , 2000 , , 364 , 732 dyer , k. k. , reynolds , s. p. , borkowski , k. j. , allen , g. e. & petre , r. 2001 , , 551 , 439 dyer , k. k. , reynolds , s. p. & borkowski , k. j. 2004 , , 600 , 752 ellison , d. c. & reynolds , s. p. 1991 , , 382 , 242 ellison , d. c. , slane , p. o. & gaensler , b. m. 2001 , , 563 , 191 ellison , d. c. & vladimirov , a. 2008 , , 673 , l47 enomoto , r. , tanimori , t. , naito , t. , yoshida , t. , yanagita , s. , mori , m. , edwards , p. g. , asahara , a. , bicknell , g. v. , gunji , s. , hara , s. , hara , t. et al . , 2002 , nature , 416 , 823 fang , j. , zhang , l. , zhang , j. f. , tang , y.y . & yu , h. 2009 , , 392 , 925 filipovi , m. d. , jones , p. a. & aschenbach , b. 2001 , young supernova remnants : eleventh astrophysics conference , " aip conference proceedings , volume 565 , pp . 267 fruscione , a. et al . 2006 , proc . spie , 6270 , 60 fukui , y. , moriguchi , y. , tamira , k. , yamamoto , h. , tawara , y. , mizuno , n. , onishi , t. , mizuno , a. , uchiyama , y. , hiraga , j. , takahashi , t. , yamashita , k. & ikeuchi , s. 2003 , , 55 , l61 garmire , g. p. , bautz , m. w. , ford , p. g. , nousek . j. a. & ricker , jr . , g. r. 2003 , spie , 4851 , 28 gooch , r. , 2006 , karma users manual , atnf gotthelf , e. v. , koralesky , b. , rudnick , l. , jones , t. w. , hwang , u. & petre , r. 2001 , , 552 , l39 green , d. a. & stephenson , f. r. 2004 , astroparticle physics , 20 , 613 green , d. a. , 2009a , bulletin of the astronomical society of india , 37 , 45 . green , d. a. , 2009b , ` a catalogue of galactic supernova remnants ( 2009 march version ) ' , astrophysics group , cavendish laboratory , cambridge , united kingdom ( available at http://www.mrao.cam.ac.uk/surveys/snrs/ " ) . griffith , m. r. & wright , a. e. 1993 , , 105 , 1666 hara , t. et al . 1993 , nucl . a 332 , 300 helder , e. a. & vink , j. 2008 , , 686 , 1094 hendrick , s. p. & reynolds , s. p. 2001 , , 559 , 903 hiraga , j. s. , uchiyama , y. , takahashi , t. & aharonian , f. a. 2005 , , 431 , 953 hiraga , j. s. , kobayashi , y. , tamagawa , t. , hayato , a. , bamba , a. , terada , y. , petre , r. , katagiri , h. & tsunemi , h. 2009 , , 61 , 275 holt , s. s. , gotthelf , e. v. , tsunemi , h. & negoro , h. 1994 , , 46 , l151 huang , c .- y . , park , s .- e . , pohl , m. & daniels , c. d. 2007 , astroparticle physics , 27 , 429 hughes , j. p. 2000 , , 545 , l53 hwang , u. , holt , s. s. & petre , r. 2000 , , 537 , l118 hwang , u. , decourchelle , a. , holt , s. s. & petre , r. 2002 , , 581 , 1101 iyudin , a. f. , schnfelder , v. , bennett , k. , bloemen , h. , diehl , r. , hermsen , w. , lichti , g. g. , van der meulen , r. d. , ryan , j. & winkler , c. 1998 , nature , 396 , 142 iyudin , a. f. , aschenbach , b. , becker , w. , dennerl , k. & haberl , f. 2005 , , 429 , 225 jones , t. j. , rudnick , l. , delaney , t. & bowden , j. 2003 , , 587,227 katagiri , h. , enomoto , r. , ksenofontov , l. t. , mori , m. , adachi , y. , asahara , a. , bicknell , g. v. , clay , r. w. , doi , y. , edwards , p. g. , gunji , s. , hara , s. et al . 2005 , , 619 , l163 katsuda , s. , tsunemi , h. & mori , k. , 2008 , , 678 , l35 katsuda , s. , petre , r. , long , k. s. , reynolds , s. p. , winkler , p. f. , mori , k. & tsunemi , h. 2009a , , 692 , l105 katsuda , s. , petre , r. , hwang , u. , yamaguchi , h. , mori , k. and tsunemi , h. 2009b , , 61 , s155 katz , b. & waxman , e. 2008 , , 1 , 18 kawachi , a. , hayami , y. , jimbo , j. , kamei , s. , kifune , t. , kubo , h. , kushida , j. , lebohec , s. , miyawaki , k. , mori , m. , nishijima , k. , patterson , j. r. , et al . 2001 , astroparticle physics , 14 , 261 keohane , j. w. 1998 , ph.d thesis , university of minnesota ksenofontov , l. t. , vlk , h. j. , berezhko , e. g. 2010 , , 714 , 1187 koyama , k. , petre , r. , gotthelf , e. v. , hwang , u. , matsuura , m. , ozaki , m. & holt , s. s. 1995 , nature , 378 , 255 koyama , k. , kinugasa , k. , matsuzaki , k. , nishiuchi , m. , sugizaki , m. , torii , k. , yamauchi , s. & aschenbach , b. 1997 , , 49 , l7 koyama , k. , ueno , m. , bamba , a. & ebisawa , k. 2001 , proceedings of the symposium new visions of the x - ray universe in the xmm - newton and chandra era , " noordwijk , the netherlands ( arxiv : astro - ph/0202009 ) lazendic , j. s. , slane , p. o. , gaensler , b. m. , reynolds , s. p. , plucinsky , p. p. and hughes , j. p. 2004 , , 602 , 271 long , k. s. , reynolds , s. p. , raymond , j. c. , winkler , p. f. , dyer , k. k. & petre , r. 2003 , , 586 , 1162 longair , m. s. ( 1994 ) , high energy astrophysics , volume 2 , stars , the galaxy and the interstellar medium , cambridge university press , cambridge , uk . lu , f. j. & aschenbach , b. 2000 , , 362 , 1083 lyutikov , m. & pohl , m. 2004 , , 609 . 785 maeda , y. , uchiyama , y. , bamba , a. , kosugi , h. , tsunemi , h. , helder , e. a. , vink , j. , kodaka , n. , terada , y. , fukazawa , y. , hiraga , j. , hughes , j. p. , kokubun , m. , kouzu , t. , matsumoto , h. , miyata , e. , nakamura , r. , okada , s. , someya , k. , tamagawa , t. , tamura , k. , totsuka , k. , tsuboi , y. , ezoe , y. , holt , s. s. , ishida , m. , kamae , t. , petre , r. & takahashi , t. 2008 , , 61 , 1217 malkov , m. a. , diamond , p. h. & sagdeev , r. z. 2005 , , 624 , l37 may , j. , murphy , d. c. & thaddeus , p. 1988 , , 73 , 51 mcclure - griffiths , n. m. , green , a. j. , dickey , j. m. , gaensler , b. m. , haynes , r. f. & wieringa , m. h. 2001 , , 551 , 394 mereghetti , s. 2001 , , 548 , l213 moraitis , k. & mastichiadis , a. 2007 , , 462 , 173 morlino , g. , amato , e. & blasi , p. 2009 , , 392 , 240 motizuki , y. , takahashi , k. , makishima , k. , bamba , a. , nakai , y. , yano , y. , igarashi , m. , motoyama , h. , kamiyama , k. , suzuki , k. & imamura , t. 2009 , nature , submitted ( http://arxiv.org/abs/0902.3446 ) muraishi , h. , tanimori , t. , yanagita , s. , yoshida , t. , moriya , m. , kifune , t. , dazeley , s. a. , edwards , p. g. , gunji , s. , hara , s. , hara , t. , kawachi , a. , et al . 2000 , , 354 , l57 murphy , d. c. 1985 , ph.d thesis , massachusetts institute of technology nakamura , r. , bamba , a. , ishida , m. , nakajima , h. , yamazaki , r. , terada , y. , phlhofer , g. & wagner , s. j. 2009 , , 61 , 197 ogasawara , t. , yoshida , t. , yamagita , s. & kifune , t. 2007 , , 309 , 401 pannuti , t. g. , allen , g. e. , houck , j. c. & sturner , s. j. 2003 , , 593 , 377 pannuti , t. g. & allen , g. e. 2004 , adspr , 33 , 434 park , s. , kargaltsev , o. , pavlov , g. g. , mori , k. , slane , p. o. , hughes , j. p. , burrows , d. n. and garmire , g. p. 2009 , apj , 695 , 431 pavlov , g. g. , sanwal , d. , kiziltan , b. & garmire , g. p. 2001 , , 559 , l131 pellizzoni , a. , mereghetti , s. & de luca , a. 2002 , , 393 , l65 plaga , r. 2008 , new astronomy , 13 , 73 pohl , m. , yan , h. & lazarian , a. 2005 , , 626 , l101 redman , m. p. , meaburn , j. , oconnor , j. a. , holloway , a. j. & bryce , m. 2000 , , 543 , l153 redman , m. p. , meaburn , j. , bryce , m. , harman , d. j. & obrien , t. j. 2002 , , 336 , 1093 redman , m. p. & meaburn , j. 2005 , , 356 , 969 reed , j. e. , hester , j. j. , fabian , a. c. and winkler , p. f. 1995 , , 440 , 706 reimer , o. & pohl , m. 2002 , , 390 , l43 reynolds , s. p. 1996 , , 459 , l13 reynolds , s. p. 1998 , , 493 , 375 reynolds , s. p. & keohane , j. w. 1999 , , 525 , 368 reynolds , s. p. , borkowski , k. j. , hwang , u. , hughes , j. p. , badenes , c. , laming , j. m. & blondin , j. m. 2007 , , 668 , l135 reynolds , s. p. , borkowski , k. j. , green , d. a. , hwang , u. , harrus , i. & petre , r. 2008 , , 680 , l41 reynolds , s. p. , borkowski , k. j. , green , d. a. , hwang , u. , harrus , i. & petre , r. 2009 , , 695 , l149 reynoso , e. m. & goss , w. m. 1999 , , 118 , 926 rosado , m. , ambrocio - cruz , p. , le coarer , e. & marcelin , m. 1996 , , 315 , 243 rho , j. , dyer , k. k. , borkowski , k. j. & reynolds , s.p . 2002 , , 581 , 1116 rho , j. , reynolds , s. p. , reach , w. t. , jarrett , t. h. , allen , g. e. & wilson , j. c. 2003 , , 592 , 299 rood , r. t. , sarazin , c. l. , zeller , e. j. & parker , b. c. 1979 , nature , 282 , 701 rothenflug , r. , ballet . , j. , dubner , g. , giacani , e. , decourchelle , a. & ferrando , p. 2004 , , 425 , 121 sankrit , r. , blair , w. p. & raymond , j. c. 2003 , , 589 , 242 sault , r. & killeen , n. 2006 , miriad users guide , atnf schnfelder , v. , bloemen , h. , collmar , w. , diehl , r. , hermsen , w. , kndlseder , j. , lichti , g. g. , plschke , s. , ryan , j. , strong , a. & winkler , c. 2000 , in aip conf . proc . 510 , fifth compton symposium , ed . m. l. mcconnell & j. m. ryan ( new york : aip ) , 54 slane , p. , gaensler , b. m. , dame , t. m. , hughes , j. p. , plucinsky , p. p. & green , a. 1999 , , 525 , 357 slane , p. , hughes , j. p. , edgar , r. j. , plucinsky , p. p. , miyata , e. , tsunemi , h. & aschenbach , b. 2001a , , 548 , 814 slane , p. , hughes , j. p. , edgar , r. j. , plucinsky , p. p. , miyata , e. , tsunemi , h. & aschenbach , b. 2001b , in holt , s. s. hwang , u. , eds , young supernova remnants , aip , new york , p. 403 stage , m. d. , allen , g. e. , houck , j. c. & davis , j. e. 2006 , natph , 2 , 614 stupar , m. , filipovi , m. d. , jones , p. a. & parker , q. a. 2005 , adspr , 35 , 1047 sugizaki , m. 1999 , ph.d thesis , university of tokyo takahashi , t. , tanaka , t. , uchiyama , y. , hiraga , j. s. , nakazawa , k. , watanabe , s. , bamba , a. , hughes , j. p. , katagiri , h. , takaoka , j. , kokubun , m. , koyama , k. , mori , k. , petre , r. , takahashi , h. & tsuboi , y. 2008 , , 60 , 131 tanaka , t. , uchiyama , y. , aharonian , f. a. , takahashi , t. , bamba , a. , hiraga , j. s. , kataoka , j. , kishishita , t. , kokubun , m. , mori , k. , nakazawa , k. , petre , r. et al . 2008 , , 685 , 988 tanimori , t. , hayami , y. , kamei , s. , dazeley , s. a. , edwards , p. g. , gunji , s. , hara , s. , hara , t. , holder , j. , kawachi , a. , kifune , t. , kita , r. , et al . 1998 , , 497 , l25 telezhinsky , i. 2009 , aph , 31 , 431 torii , k. , uchida , h. , hasuike , k. , tsunemi , h. , yamauchi , y. & shibata , s. , , 58 , l11 tsunemi , h. , miyata , e. , aschenbach , b. , hiraga , j. & akutsu , d. 2000 , , 52 , 887 turtle , a. j. , pugh , j. f. , kenderdine , s. , pauliny - toth , i. i. k. 1962 , , 124 , 297 uchiyama , y. , aharonian , f. a. & takahashi , t. 2003 , , 400 , 567 uchiyama , y. , aharonian , f. a. , tanaka , t. , takahashi , t. & maeda , y. 2007 , nature , 449 , 576 ueno , m. , bamba , a. , koyama , k. & ebisawa , k. 2003 , , 588 , 338 van der swaluw , e. & achterberg , a. 2004 , , 421 , 1021 vink , j. 2003 , the restless high energy universe " conference proceedings , nucl . physics b. suppl . series , eds . e. p. j. van den heuvel , j. j. m. in t zand & r. a. m. j. wijers , 132 , 21 vink , j. , laming , j. m. , gu , m. f. , rasmussen , a. & kaastra , j. s. 2003 , , 587 , l31 vink , j. & laming , j. m. 2003 , , 584 , 758 vink , j. 2005 , international symposium on high - energy gamma - ray astronomy ( gamma-2004 ) " conference proceedings , heidelberg , edited by f. a. aharonian and h. vlk ( aip , ny ) , 745 , 160 vink , j. , bleeker , j. , van der heyden , k. , bykov , a. , bamba , a. & yamazaki , r. 2006 , , 648 , l33 vink , j. 2008 , , 486 , 837 weisskopf , m. c. , brinkman , b. , canizares , c. , garmire , g. , murray , s. & van speybroeck , l. p. et al . 2002 , , 114 , 1 winkler , p. f. & long , k. s. 1997 , , 491 , 829 winkler , p. f. , gupta , g. & long , k. s. 2003 , , 585 , 324 yamaguchi , h. , ueno , m. , koyama , k. , bamba , a. & yamauchi , s. 2004 , , 56 , 1059 zirakashvili , v. n. & aharonian , f. 2007 , , 465 , 695 lcccccccc sn 1006 & 2.2 & ( 1 ) & 15 & 420 & 0.040.20 & 0.00440.022 & ( 2 ) , ( 3 ) + cas a & 3.4 & ( 4 ) & 2.55 & 3 - 4 & 0.05 & 0.02 - 0.03 & ( 5 ) , ( 6 ) + kepler & 4.8 & ( 7 ) & 1.5 & 2 - 3 & 0.050.07 & 0.020.03 & ( 8) + tycho & 2.3 & ( 9 ) & 4 & 5 & 0.06 & 0.021 & ( 8) , ( 10 ) + rcw 86 & 2.8 & ( 11 ) & 21 & 20 & 0.27 & 0.016 & ( 8) + [email protected] & 5.0 & ( 12 ) & 5.5 & 1216 & 0.30.4 & 0.0360.048 & ( 13 ) + [email protected] & 6.3 & ( 14 ) & 30 & 2040 & 0.61.2 & 0.0110.022 & ( 15 ) , ( 16 ) + & 1.0 & ( 17 ) & 30 & 2040 & 0.10.2 + cccccc a & 08 49 19.44 & @xmath80 34 51.2 & @xmath81 & @xmath82 & 3.4 + b & 08 49 07.37 & @xmath80 35 52.0 & @xmath83 & @xmath84 & 1.9 + c & 08 49 03.99 & @xmath80 36 16.9 & @xmath85 & @xmath86 & 1.7 + d & 08 48 57.37 & @xmath80 37 15.6 & @xmath87 & @xmath88 & 2.2 + e & 08 48 52.70 & @xmath80 37 50.9 & @xmath89 & @xmath90 & 2.9 + f & 08 48 47.33 & @xmath80 38 27.8 & @xmath91 & @xmath92 & 3.9 + g & 08 48 42.69 & @xmath80 39 02.7 & @xmath93 & @xmath94 & 4.9 + h & 08 48 38.77 & @xmath80 39 52.3 & @xmath95 & @xmath96 & 5.8 + lcccccc sn 1006 & # 1 & 1.7 ( 1.3 , 2.2 ) & 0.02 ( 0.01 , 0.02 ) & 160 ( @xmath9792 ) & 1.7 ( @xmath970.98 ) & ( 1 ) + & # 2 & 14 ( 9.8 , 21 ) & 0.15 ( 0.10 , 0.22 ) & 93 ( 73 , 130 ) & 0.99 ( 0.78 , 1.39 ) + & # 3 & 2.3 ( 0.78 , 4.3 ) & 0.02 ( 0.01 , 0.05 ) & 150 ( @xmath9756 ) & 1.60 ( @xmath970.60 ) + & # 4 & 0.12 ( ) & 0.001 ( ) & 150 ( @xmath9778 ) & 1.60 ( @xmath970.83 ) + & # 5 & 4.5 ( 2.2 , 8.7 ) & 0.05 ( 0.02 , 0.09 ) & 62 ( 43 , 100 ) & 0.66 ( 0.46 , 1.07 ) + & # 6 & 3.4 ( 2.4 , 5.0 ) & 0.04 ( 0.03 , 0.05 ) & 130 ( 86 , 250 ) & 1.39 ( 0.92 , 2.67 ) + cas a & # 1 & @xmath980.93 & @xmath980.02 & 1.27 ( 0.96 , 1.81 ) & 0.02 ( 0.02 , 0.03 ) & ( 2 ) + & # 2 & @xmath980.80 & @xmath980.01 & @xmath981.59 & @xmath980.03 + kepler & # 1 & 1.17 ( 0.87 , 1.59 ) & 0.03 ( 0.02 , 0.04 ) & 0.93 ( @xmath981.41 ) & 0.02 ( @xmath980.03 ) & ( 2 ) + & # 2 & 1.59 ( 1.17 , 2.89 ) & 0.04 ( 0.03 , 0.05 ) & 3.09 ( 3.46 , 3.87 ) & 0.07 ( 0.08 , 0.09 ) + tycho & # 1 & 1.18 ( 1.01 , 1.32 ) & 0.01 ( 0.01 , 0.01 ) & 5.36 ( 4.47 , 6.12 ) & 0.06 ( 0.05 , 0.07 ) & ( 2 ) + & # 2 & @xmath980.80 & @xmath980.009 & 1.70 ( 1.32 , 3.15 ) & 0.02 ( 0.01 , 0.04 ) + & # 3 & @xmath980.80 & @xmath980.009 & 2.38 ( 2.20 , 2.54 ) & 0.03 ( 0.02 , 0.03 ) + & # 4 & 0.86 ( 0.80 , 0.93 ) & 0.01 ( 0.009 , 0.01 ) & 5.53 ( 5.00 , 6.14 ) & 0.06 ( 0.06 , 0.07 ) + & # 5 & 1.03 ( 0.90 , 1.35 ) & 0.01 ( 0.01 , 0.02 ) & 2.47 ( 1.93 , 3.15 ) & 0.03 ( 0.02 , 0.04 ) + rcw 86 & # 1 & 2.39 ( 1.48 , 3.34 ) & 0.03 ( 0.02 , 0.05 ) & 20.1 ( 17.3 , 23.8 ) & 0.27 ( 0.23 , 0.32 ) & ( 2 ) + & # 2 & 1.56 ( 0.49 , 4.79 ) & 0.02 ( 0.007 , 0.07 ) & 18.2 ( 11.8 , 35.6 ) & 0.25 ( 0.16 , 0.48 ) + [email protected] & 1 & 19.0 ( 12.1 , 31.4 ) & 0.07 ( 0.04 , 0.11 ) & 31.8 ( 23.7 , 47.1 ) & 0.12 ( 0.09 , 0.17 ) & ( 3 ) + & 2 & 3.68 ( 2.69 , 5.75 ) & 0.01 ( 0.01 , 0.02 ) & 65.0 ( 38.2 , 144.6 ) & 0.24 ( 0.14 , 0.53 ) + & 3 & & & 37.1 ( 26.5 , 60.2 ) & 0.13 ( 0.10 , 0.22 ) + [email protected] & a & 16.24 ( 11.53 , 18.15 ) & 0.06 ( 0.04 , 0.07 ) & & & ( 4 ) + & b & 7.04 ( 6.29 , 8.25 ) & 0.03 ( 0.02 , 0.03 ) & & + & c & 7.60 ( 6.68 , 8.66 ) & 0.03 ( 0.02 , 0.03 ) & & + & d & 10.83 ( 9.78 , 12.23 ) & 0.04 ( 0.03 , 0.04 ) & & + & e & 10.80 ( 9.51 , 11.57 ) & 0.04 ( 0.03 , 0.04 ) & & + & f & 5.88 ( 5.37 , 6.41 ) & 0.02 ( 0.02 , 0.02 ) & & + & g & 17.17 ( 16.15 , 17.95 ) & 0.06 ( 0.06 , 0.07 ) & & + & h & 21.67 ( 20.02 , 23.55 ) & 0.08 ( 0.07 , 0.09 ) & & + lcccccc # 1 & s2 & 08 49 24 & @xmath045 34 52 & 4.95 & leading rim & 0.036 + # 2 & s2 & 08 49 32 & @xmath045 36 48 & 4.39 & trailing rim & 0.032 + # 3 & s3 & 08 48 53 & @xmath045 38 05 & 5.15 & leading shock & 0.037 + # 4 & s3 & 08 48 57 & @xmath045 38 31 & 10.20 & leading rim & 0.074 + # 5 & s3 & 08 49 02 & @xmath045 40 34 & 18.10 & interior region & 0.131 + # 6 & s3 & 08 49 17 & @xmath045 39 34 & 3.21 & trailing rim & 0.023 + # 7 & s4 & 08 48 24 & @xmath045 42 57 & 5.15 & leading shock & 0.037 + lcccccccccc # 1 & 0.43@xmath99 & 2.68@xmath100 & 6.58@xmath101 & 6.48@xmath101 & 270.20/279=0.97 & 7.63@xmath910@xmath102 & 1.04@xmath910@xmath103 & 7.52@xmath910@xmath102 & 1.02@xmath910@xmath103 + # 2 & 0.35@xmath104 & [email protected] & 4.94@xmath105 & 4.94@xmath106 & 238.15/263=0.91 & 6.60@xmath910@xmath102 & 8.40@xmath910@xmath102 & 6.60@xmath910@xmath102 & 8.40@xmath910@xmath102 + # 3 & [email protected] & [email protected] & 7.36@xmath107 & 7.58@xmath108 & 320.67/343=0.93 & 1.09@xmath910@xmath103 & 1.42@xmath910@xmath103 & 1.12@xmath910@xmath103 & 1.43@xmath910@xmath103 + # 4 & [email protected] & 2.52@xmath109 & 20.20@xmath110 & 20.00@xmath110 & 519.06/477=1.09 & 2.67@xmath910@xmath103 & 3.51@xmath910@xmath103 & 2.65@xmath910@xmath103 & 3.48@xmath910@xmath103 + # 5 & [email protected] & 2.60@xmath100 & 7.99@xmath111 & 8.12@xmath112 & 406.38/393=1.03 & 1.06@xmath910@xmath103 & 1.36@xmath910@xmath103 & 1.08@xmath910@xmath103 & 1.38@xmath910@xmath103 + # 6 & 0.39@xmath113 & 2.70@xmath114 & 2.25@xmath115 & 2.15@xmath116 & 203.13/194=1.05 & 2.61@xmath910@xmath102 & 3.47@xmath910@xmath102 & 2.50@xmath910@xmath102 & 3.32@xmath910@xmath102 + # 7 & [email protected] & [email protected] & 5.41@xmath117 & 5.22@xmath118 & 310.23/277=1.12 & 8.56@xmath910@xmath102 & 1.04@xmath910@xmath103 & 8.26@xmath910@xmath102 & 1.00@xmath910@xmath103 lccccccc # 1 & 0.18@xmath119 & 0.54@xmath120 & 1.65@xmath121 & 0.036 & 282.55/280=1.01 & 7.75@xmath910@xmath102 & 8.70@xmath910@xmath102 + # 2 & 0.17@xmath122 & 0.55@xmath123 & 1.83@xmath124 & 0.032 & 245.52/264=0.93 & 6.68@xmath910@xmath102 & 7.46@xmath910@xmath102 + # 3 & 0.32@xmath125 & 0.52@xmath126 & [email protected] & 0.037 & 325.13/344=0.95 & 1.11@xmath910@xmath103 & 1.34@xmath910@xmath103 + # 4 & 0.24@xmath127 & 0.52@xmath123 & [email protected] & 0.074 & 549.07/478=1.15 & 2.70@xmath910@xmath103 & 3.13@xmath910@xmath103 + # 5 & 0.21@xmath128 & 0.59@xmath129 & 1.93@xmath124 & 0.131 & 410.81/394=1.04 & 1.09@xmath910@xmath103 & 1.24@xmath910@xmath103 + # 6 & 0.17@xmath130 & 0.58@xmath131 & [email protected] & 0.023 & 212.45/195=1.09 & 2.63@xmath910@xmath102 & 2.93@xmath910@xmath102 + # 7 & 0.21@xmath132 & 0.54@xmath133 & [email protected] & 0.037 & 314.72/278=1.13 & 8.42@xmath910@xmath102 & 9.64@xmath910@xmath102 lccccccc # 1 & 0.39@xmath134 & 0.50@xmath135 & 2.47@xmath136 & 0.036 & 273.06/280=0.98 & 7.51@xmath910@xmath102 & 9.90@xmath910@xmath102 + # 2 & 0.37@xmath137 & 0.51@xmath138 & 2.59@xmath139 & 0.032 & 245.25/264=0.93 & 6.48@xmath910@xmath102 & 8.39@xmath910@xmath102 + # 3 & 0.51@xmath140 & 0.49@xmath141 & 2.63@xmath142 & 0.037 & 362.21/344=1.05 & 1.06@xmath910@xmath103 & 1.50@xmath910@xmath103 + # 4 & 0.41@xmath143 & 0.48@xmath144 & 2.86@xmath145 & 0.074 & 540.18/478=1.13 & 2.61@xmath910@xmath103 & 3.46@xmath910@xmath103 + # 5 & 0.38@xmath146 & 0.56@xmath138 & 2.79@xmath147 & 0.131 & 413.90/394=1.05 & 1.04@xmath910@xmath103 & 1.37@xmath910@xmath103 + # 6 & 0.39@xmath148 & 0.52@xmath149 & 1.88@xmath150 & 0.023 & 205.49/195=1.05 & 2.50@xmath910@xmath102 & 3.34@xmath910@xmath102 + # 7 & 0.42@xmath151 & 0.50@xmath152 & 2.47@xmath153 & 0.037 & 337.09/278=1.21 & 8.10@xmath910@xmath102 & 1.07@xmath910@xmath103 + lccc ctb 37b & 3.6@xmath910@xmath154 & 150 & ( 1 ) + kes 73 & 1.5@xmath910@xmath154 & 100 & ( 2 ) + [email protected] & 1.4@xmath910@xmath154 & 90 & ( 3 ) + cas a & 9.0@xmath910@xmath155 & 80 & ( 4 ) + & 3.2@xmath910@xmath155 & 50 & ( 2 ) + [email protected] & 7.0@xmath910@xmath155 & 70 & ( 4 ) + [email protected] & 3.3@xmath910@xmath155 & 50 & ( 6 ) + [email protected] & 2.8@xmath910@xmath155 & 40 & ( 7 ) + [email protected] & 2.4@xmath910@xmath155 & 40 & ( 8) + & 1.5@xmath910@xmath156 & 9.7 & ( 9 ) + [email protected] & 6.3@xmath910@xmath155 & 60 & ( 10 ) + & 2.3@xmath910@xmath155 & 40 & ( 11 ) + & 2.2@xmath910@xmath155 & 40 & ( 12 ) + sn 1006 & 1.1@xmath910@xmath155 & 30 & ( 13 ) + & 6.0@xmath910@xmath156 & 20 & ( 14 ) + kepler & 1.1@xmath910@xmath155 & 30 & ( 2 ) + tycho & 8.8@xmath910@xmath156 & 20 & ( 2 ) + rcw 86 & 8.5@xmath910@xmath156 & 20 & ( 15 ) + [email protected] & 6.6@xmath910@xmath156 & 20 & ( 2 ) + 3c 397 & 3.4@xmath910@xmath156 & 20 & ( 2 ) + w49b & 2.4@xmath910@xmath156 & 10 & ( 2 ) + [email protected] & 1.8@xmath910@xmath156 & 10 & ( 2 ) + 3c 396 & 1.6@xmath910@xmath156 & 10 & ( 2 ) + [email protected] & 1.5@xmath910@xmath156 & 9.7 & ( 2 ) + 3c 391 & 1.4@xmath910@xmath156 & 9.4 & ( 2 ) + [email protected] ( sn 386 ) & 1.2@xmath910@xmath156 & 8.7 & ( 2 ) + rcw 103 & 1.2@xmath910@xmath156 & 8.7 & ( 2 ) lcc dem l71 & 2.49@xmath910@xmath155 & 40 + n49b & 6.76@xmath910@xmath156 & 20 + n103b & 4.35@xmath910@xmath156 & 20 + n23 & 4.11@xmath910@xmath156 & 20 + n132d & 3.14@xmath910@xmath156 & 10 + [email protected] & 2.90@xmath910@xmath156 & 10 + [email protected] & 1.93@xmath910@xmath156 & 10 + n63a & 1.93@xmath910@xmath156 & 10 + [email protected] & 2.42@xmath910@xmath157 & 3.9 + [email protected] & 1.45@xmath910@xmath157 & 3.0 + [email protected] & 9.66@xmath910@xmath158 & 2.5	we present a detailed spatially - resolved spectroscopic analysis of two x - ray observations ( with a total integration time of 73280 seconds ) made of the luminous northwestern rim complex of the galactic supernova remnant ( snr ) [email protected] ( rx j0852.0@xmath04622 ) with the _ chandra _ x - ray observatory . [email protected] is a member of a class of galactic snrs which feature x - ray spectra dominated by non - thermal emission : in the cases of these snrs , the emission is believed to have a synchrotron origin and studies of the x - ray spectra of these snrs can lend insights into how snrs accelerate cosmic - ray particles . the _ chandra _ observations have clearly revealed fine structure in this rim complex ( including a remarkably well - defined leading shock ) and the spectra of these features are dominated by non - thermal emission . we have measured the length scales of the upstream structures at eight positions along the rim and derive lengths of 0.02 - 0.08 pc ( assuming a distance of 750 pc to [email protected] ) . we have also extracted spectra from seven regions in the rim complex ( as sampled by the acis - s2 , -s3 and -s4 chips ) and fit these spectra with such models as a simple power law as well as the synchrotron models srcut and sresc . we have constrained our fits to the latter two models using estimates for the flux densities of these filaments at 1 ghz as determined from radio observations of this rim complex made with the australia telescope compact array ( atca ) . statistically - acceptable fits to all seven regions are derived using each model : differences in the fit parameters ( such as photon index and cutoff frequency ) are seen in the different regions , which may indicate variations in shock conditions and the maximum energies of the cosmic - ray electrons accelerated at each region . finally , we estimate the maximum energy of cosmic - ray electrons accelerated along this rim complex to be approximately 40 tev ( corresponding to one of the regions of the leading shock structure assuming a magnetic field strength of 10 @xmath1 g ) . we include a summary of estimated maximum energies for both galactic snrs as well as snrs located in the large magellanic cloud . like these other snrs , it does not appear that [email protected] is currently accelerating electrons to the knee energy ( @xmath23000 tev ) of the cosmic - ray spectrum . this result is not surprising , as there is some evidence that loss mechanisms which are not important for the accelerated cosmic - ray nucleons at energies just below the knee might cut off electron acceleration .
recent high - resolution direct imaging observations have revealed spiral structure in three protoplanetary disks around herbig ae / be stars : sao 206462 ( muto et al . 2012 ; garufi et al . 2013 ) , mwc 758 ( grady et al . 2013 ; benisty et al . 2015 ) , and hd 100546 ( currie et al . the polarized intensity has been measured in these observations to gain higher contrast between the disk and the central star . while the thermal emission from the central star is unpolarized , the scattered light from the disk is polarized . in these near - infrared ( near - ir ) polarized intensity images , two spiral arms with roughly 180@xmath0 rotational symmetry are present in both sao 206462 and mwc 758 , similar to the grand design in a spiral galaxy ( e.g. the whirlpool galaxy m51 ) . the spiral arms also exhibit a high contrast against the background disk . the polarized intensity of the spiral arm is several times higher than that of the region outside the spiral arm . it should also be noted that , since the dust scattering opacity is quite large , these observations only probe structure high up at the disk atmosphere ( e.g. several disk scale heights ) where the last dust scattering surface is . in addition to spiral patterns , these three disks also have gaps or holes ( sao 206462 with a submillimeter cavity of 46 au , mwc 758 with a cavity of 73 au from andrews et al . 2011 , and hd 100546 with a cavity of 10 au revealed by sed fitting from bouwman et al . 2003 ) , which indicates that they are members of the protoplanetary disk class called transitional disks ( espaillat et al . one scenario to explain both spiral patterns and gaps / holes is that these disks harbor low - mass companions ( e.g. young planets ) which can open gaps and excite spiral waves at the same time ( e.g. , baruteau et al . 2014 ) . however , there are two difficulties in explaining the observed spiral patterns using planet - induced spiral wakes . first , the large pitch angle of all the observed spiral arms suggests that the disk has a relative high temperature ( e.g. @xmath1250 k at 70 au for mwc 758 , benisty et al . 2015 ) . in linear theory , spiral waves are basically sound waves in disks , and the pitch angle of the spiral arms is directly related to the sound speed in the disk . using the linear theory , the best fit models for both sao 206462 ( muto et al . 2012 ) and mwc 758 ( grady et al . 2013 ; benisty et al . 2015 ) suggest that the disk aspect ratio ( @xmath2 with @xmath3 ) at @xmath4 au is around 0.2 which is too large for any realistic disk structure . for example , even if the stellar irradiation is perpendicular to the disk surface , the maximum disk temperature due to the stellar irradiation is @xmath5 so that @xmath6 . assuming a 2 @xmath7 central star with 10 @xmath8 luminosity , the maximum temperature is @xmath170 k at 100 au and @xmath2 is only @xmath10.1 . since @xmath9 , it is very difficult to make @xmath100.2 . second , the observed spiral arms exhibit much higher brightness contrasts than suggested by the synthetic observations based on two dimensional ( 2-d ) planet - disk simulations . juhsz et al . ( 2015 ) have calculated the polarized scattered light images by combining 2-d hydrodynamical simulations with 3-d monte - carlo radiative transfer ( mcrt ) simulations . vertical hydrostatic equilibrium has been assumed to extend the 2-d simulation to the third dimension ( the vertical direction ) . they find that a relative change of about 3.5 on the spiral arms in the surface density is required for the spirals to be detectable . this value is a factor of eight higher than what is seen in their hydrodynamical simulations . in this paper , we first point out that the pitch angle formula derived from the linear theory , which has been used in almost all previous spiral arm modeling efforts , does not apply to the high planet mass cases . spiral wakes that are excited by high mass planets ( e.g. 1 @xmath11 ) become spiral shocks which propagate at speeds faster than the local sound speed ( goodman & rafikov 2001 , rafikov 2002 ) . thus , the pitch angle difficulty above can be alleviated by considering the non - linear extension of the spiral shock theory . we also show that spiral arms ( especially the inner arms ) have complicated non - hydrostatic 3-d structure . such structure can lead to strong density perturbation at the disk surface resulting in a corrugated shape of its atmosphere . since near - ir observations only probe the shape of disk surface , this effect alleviates the second difficulty mentioned above . in dong et al . ( 2015 ) , we have combined mcrt simulations with hydrodynamical simulations from this paper and demonstrated that planet - induced inner spiral arms can explain recent near - ir direct imaging observations of sao 206462 and mwc 758 . we note that , since the planets that we have proposed are outside the spiral arms , we can not explain the gaps discovered at small radii in these transitional disks . other mechanisms , e.g. another planet or photoevaporation , are needed to explain these gaps . before we introduce our numerical method in 3 , we provide the theoretical background in 2 . the shape of the spiral wakes will be studied in 4 , and their 3-d structure will be presented in 5 . after a short discussion in 6 , we summarize our results in 7 . as a result of planet - disk interaction , a spiral arm forms due to the constructive interference of density wakes with different azimuthal wavenumbers @xmath12 excited by the planet at lindblad resonances . in the linear density wave theory , the @xmath12-th fourier component of the planet potential excites the density wave having @xmath12 spiral arms @xmath13}\ ] ] where @xmath14 is any perturbed quantity associated with the wave , @xmath15 is its complex amplitude , @xmath16 is the radial wave vector , and @xmath17 is the planet orbital frequency . thus , the wave has the same phase along the curve satisfying @xmath18 . the pitch angle ( @xmath19 ) of the equal phase curve satisfies tan@xmath20 , so @xmath19=tan@xmath21\|$ ] . using the dispersion relationship for density waves in the large @xmath12 limit and far from the launching point , @xmath22 , we have @xmath19=tan@xmath23 $ ] . because @xmath19 is independent of @xmath12 , different @xmath12 modes can constructively interfere to form the one armed spiral wake ( ogilvie & lubow 2002 ) . if the equal phase curve is integrated from the planet s position ( @xmath24 , @xmath25 ) , the shape of the wake far from @xmath24 is given by rafikov ( 2002 ) and muto et al . ( 2012 ) as @xmath26\label{eq : linearwake}\end{aligned}\ ] ] where @xmath27 is the disk aspect ratio at @xmath24 , @xmath28 , and the sound speed @xmath29 . however , when the planet is massive enough , the above linear density wave theory breaks down . linear waves excited by planets will steepen to shocks ( goodman & rafikov 2001 , rafikov 2002 , dong , rafikov & stone 2012 , duffell & macfadyen 2012 , zhu et al . 2013 ) after they propagate over a distance @xmath30 where @xmath31 is the adiabatic index , and @xmath32 is the disk thermal mass @xmath33 when @xmath34 , the spiral waves will immediately become spiral shocks after they are excited around the planet . unlike the linear wake which follows equation ( [ eq : linearwake ] ) , the spiral shock will expand away from the trajectory predicted by equation ( [ eq : linearwake ] ) thus , if there is a massive planet in the disk , using equation ( [ eq : linearwake ] ) to fit the shape of the spiral shocks will predict an incorrect disk aspect ratio and temperature . to study density wakes / shocks excited by planets , we have carried out both 2-d and 3-d hydrodynamical simulations using athena++ . athena++ is a newly developed grid based code using a higher - order godunov scheme for mhd and the constrained transport ( ct ) to conserve the divergence - free property for magnetic fields ( stone et al . , in preparation ) . but in this paper , we do not include magnetic fields and only solve hydrodynamical equations using athena++ . compared with its predecessor athena ( gardiner & stone 2005 , 2008 ; stone et al . 2008 ) , athena++ is highly optimized and uses flexible grid structures , allowing global numerical simulations spanning a large radial range . furthermore , the geometric source terms in curvilinear coordinates ( e.g. in cylindrical and spherical - polar coordinates ) are carefully implemented so that angular momentum is conserved exactly ( to machine precision ) , which makes the code ideal for global disk simulations . our simulations use the adiabatic equation of state ( eos ) with the adiabatic index @xmath31=1.4 . a simple orbital cooling scheme has been applied to mimic the radiative cooling process in disks . in 3-d simulations , we have adopted @xmath35 where @xmath36 and @xmath37 are the density and the internal energy per unit volume , while in 2-d simulations , we have adopted @xmath38 where @xmath39 and @xmath37 are the disk surface density and the internal energy per unit area . @xmath40 is the heat capacity per unit mass , @xmath41 is the boltzmann constant , @xmath42 is the mean molecular weight , and @xmath43 is the atomic mass unit . the cooling time @xmath44 can be written in the dimensionless form as @xmath45 . we fix @xmath46 to be a constant in each simulation . with this scheme , the disk temperature is relaxed to the background disk temperature ( @xmath47 ) determined by stellar irradiation . in our simulations , @xmath47 is set to be the initial disk temperature . to estimate @xmath46 in a realistic disk , we use the grey atmosphere approximation ( hubeny 1990 ) for the radiative cooling , @xmath48 where @xmath49 is the stefan - boltzmann constant , @xmath50 is the optical depth in the vertical direction , @xmath51 is the rosseland mean opacity , and @xmath52 is the midplane temperature . assuming @xmath53 and using equations ( [ eq : dedt ] ) and ( [ eq : dedt2 ] ) , we can derive @xmath54 approximating the polynomial of @xmath52 and @xmath47 in the denominator of equation [ eq : tcool ] with @xmath55 ^ 3 $ ] and assuming the central star is 1 @xmath7 , we have @xmath56 ^ 3}\nonumber\\ & & \times \frac{1+\tau^2}{\tau}\label{eq : tcoolvalue}\end{aligned}\ ] ] thus , @xmath46 can vary dramatically at different radii in disks . using the minimum mass solar nebular model , @xmath46 is @xmath110@xmath57 at 1 au and @xmath58 at 100 au . we have carried out three sets of simulations with @xmath59 , 1 , and 100 . they are respectively labeled as iso , t1 , t2 at the end of their names in table 1 . simulations with fast cooling ( @xmath59 ) are equivalent to locally isothermal simulations . simulations with @xmath60 are basically adiabatic simulations considering the timescale of the simulations is several tens of orbits , and simulations with @xmath61 should be between isothermal and adiabatic simulations . somewhat surprisingly , we find that simulations with @xmath62 are qualitatively similar to those with @xmath60 . this similarity suggests that the spiral waves in disks with @xmath62 behave adiabatically . we think that this is due to the short timescale for the flow to move across the spiral wake . the spiral wake has a typical width smaller than the disk scale height , and the background flow moves across the wake at nearly keplerian speed . thus , when the fluid travels in and out of the wave / shock , its response time is much smaller than the orbital time and should behave adiabatically even in disks with @xmath62 . considering this similarity , in most part of the paper , we only show results with @xmath59 and 1 . cccccccc 2-d & & & & & & + run & @xmath63 & @xmath46 & domain & resolution & + & @xmath11 & & @xmath64 & @xmath65 + cm1iso & 0.01 & 10@xmath66 & [ 0.2,10 ] & 1280@xmath672048 + cm1t1 & 0.01 & 1 & [ 0.2,10 ] & 1280@xmath672048 + cm1t100 & 0.01 & 100 & [ 0.2,10 ] & 1280@xmath672048 + cm2iso & 1 & 10@xmath66 & [ 0.2,10 ] & 1280@xmath672048 + cm2t1 & 1 & 1 & [ 0.2,10 ] & 1280@xmath672048 + cm2t100 & 1 & 100 & [ 0.2,10 ] & 1280@xmath672048 + cm3iso & 6 & 10@xmath66 & [ 0.2,10 ] & 1280@xmath672048 + cm3t1 & 6 & 1 & [ 0.2,10 ] & 1280@xmath672048 + cm3t100 & 6 & 100 & [ 0.2,10 ] & 1280@xmath672048 + 3-d & & & & & & + run & @xmath63 & @xmath46 & domain & resolution & + & @xmath11 & & @xmath68 & @xmath69 + sthin & 0.000316 & 10@xmath66 & [ 0.5,2]@xmath67[@xmath70-@xmath71,@xmath70+@xmath71 ] & 456@xmath67128@xmath672048 + sm1iso & 0.01 & 10@xmath66 & [ 0.3,3]@xmath67[@xmath70-@xmath73,@xmath70+@xmath73 ] & 256@xmath67128@xmath67688 + sm1t1 & 0.01 & 1 & [ 0.3,3]@xmath67[@xmath70-@xmath73,@xmath70+@xmath73 ] & 256@xmath67128@xmath67688 + sm1t100 & 0.01 & 100 & [ 0.3,3]@xmath67[@xmath70-@xmath73,@xmath70+@xmath73 ] & 256@xmath67128@xmath67688 + sm2iso & 1 & 10@xmath66 & [ 0.3,3]@xmath67[@xmath70-@xmath73,@xmath70+@xmath73 ] & 256@xmath67128@xmath67688 + sm2t1 & 1 & 1 & [ 0.3,3]@xmath67[@xmath70-@xmath73,@xmath70+@xmath73 ] & 256@xmath67128@xmath67688 + sm2t100 & 1 & 100 & [ 0.3,3]@xmath67[@xmath70-@xmath73,@xmath70+@xmath73 ] & 256@xmath67128@xmath67688 + sm3iso & 6 & 10@xmath66 & [ 0.3,3]@xmath67[@xmath70-@xmath73,@xmath70+@xmath73 ] & 256@xmath67128@xmath67688 + sm3t1 & 6 & 1 & [ 0.3,3]@xmath67[@xmath70-@xmath73,@xmath70+@xmath73 ] & 256@xmath67128@xmath67688 + sm3t100 & 6 & 100 & [ 0.3,3]@xmath67[@xmath70-@xmath73,@xmath70+@xmath73 ] & 256@xmath67128@xmath67688 + sm2isolong & 1 & 10@xmath66 & [ 0.3,3]@xmath67[@xmath70-@xmath73,@xmath70+@xmath73 ] & 256@xmath67128@xmath67688 + sm2isocyl & 1 & 10@xmath66 & [ 0.3,3]@xmath67[-0.8,0.8 ] & 296@xmath67176@xmath67690 + sm2isoconstt & 1 & 10@xmath66 & [ 0.3,3]@xmath67[@xmath70-@xmath73,@xmath70+@xmath73 ] & 256@xmath67128@xmath67688 + we have also varied the planet mass to be 0.01 , 1 , and 6 @xmath11 in our main set of simulations , which are labeled as m1 , m2 , and m3 in their names respectively . here , we have defined @xmath11 as 0.001 of the central star s mass . the thermal mass ( equation [ eq : thermal ] ) for the @xmath75 disk is @xmath76 . thus , waves excited by a 0.01 @xmath11 planet are in the linear regime and waves from a 6 @xmath11 planet are in the highly non - linear regime . to compare with figure 2 in tanaka et al . ( 2002 ) , we have also carried out a thin disk simulation with @xmath77 ( sthin in table 1 ) . the thermal mass for such a thin disk is only @xmath780.0316 @xmath11 . thus , in order to ensure that the waves are in the linear regime , we choose the planet mass of @xmath79 in this thin disk simulation . to avoid the divergence of planet potential , a smoothing length of 0.1 @xmath24 has been applied for m2 and m3 cases ( 0.1 @xmath24 is close to the disk scale height and the hill radius of the planet ) . for the thin disk case which has a very small mass planet , we choose a smoothing length of 6@xmath6710@xmath80 , roughly the length of two grid cells . for the low mass planet cases ( m1 ) , a smoothing length of 0.02 @xmath24 , which is also roughly the length of two grid cells in these simulations , has been adopted . planets are fixed in circular orbits at @xmath81 , and the indirect potential , which is due to the center of the coordinate system is at the star instead of the center of the mass , has been included . we have run the simulations for 10 planetary orbits . we choose this timescale because it is longer than the sound crossing time throughout the whole disk so that density waves / shocks have established , while it is shorter than the gap opening timescale to avoid complicated gap structures ( e.g. vortices at the gap edges ) and other longterm effects ( e.g. radial buoyancy waves , richert et al . 2015 ) . to verify that the revealed wave mechanics still hold in long terms we have run one simulation for 120 orbits ( sm2isolong in table 1 ) , which will be discussed in 6.1 . a constant @xmath82 viscosity with @xmath83 has been applied in our main sets of simulations . at inner and outer boundaries , all quantities are fixed at the initial states . for a numerical code using godunov scheme which calculates the flux by decomposing wave characteristics at the left and right grid cells , such boundary can absorb wave characteristics coming from the active zones and limit waves traveling from the ghost zones to active zones . such boundary condition shows little wave reflection and is similar to the non - reflecting boundary condition ( godon 1996 ) used in the fargo code ( masset 2000 ) . the detailed code comparison is given in appendix c of zhu et al . 2014 . compared with the 3-d simulations in the next subsection , 2-d simulations allow us to study density wakes in a bigger domain using a higher numerical resolution . the initial radial profile of the disk is @xmath84 we choose @xmath85 , @xmath86 , and @xmath87 to make @xmath88 . cylindrical coordinates have been adopted . to make every grid cell have equal length in the radial and azimuthal direction throughout the whole domain , the grids are uniformly spaced in log(@xmath64 ) from @xmath890.2 to 10 , and uniformly spaced from @xmath90 to @xmath91 in the @xmath92 direction . our standard resolution is 1280 in the @xmath64 direction and @xmath93 in the @xmath92 direction , which is equivalent to 32 grids per @xmath94 at @xmath81 in both directions . in table 1 , 2-d runs are denoted with a `` c '' in front of the model names , while 3-d runs are denoted with a `` s '' in front of the model names . to study the 3-d structure of density wakes / shocks , we have run 3-d hydrodynamical simulations in spherical polar coordinates except for one case in cylindrical coordinates . the initial density profile of the disk at the disk midplane is @xmath95 and the temperature is constant on cylinders @xmath96 we want to emphasize that @xmath64 should not be confused with @xmath97 . in this paper , we use ( @xmath64 , @xmath92 , @xmath98 ) to represent positions in cylindrical coordinates while using ( @xmath97 , @xmath99 , @xmath92 ) for spherical polar coordinates . @xmath92 represents the azimuthal direction ( the direction of disk rotation ) in both coordinate systems . considering the disk structure is more natural to be described in cylindrical coordinates , we have transformed 3-d simulation results from spherical polar coordinates to cylindrical coordinates . most results presented below are plotted in cylindrical coordinates with @xmath64 representing the distance to the axis of the disk , even though most simulations are carried out in spherical polar coordinates . hydrostatic equilibrium in the @xmath100 plane requires that ( e.g. nelson et al . 2013 ) @xmath101\,,\ ] ] and @xmath102^{1/2}\,,\ ] ] where @xmath103 is the isothermal sound speed at @xmath64 , @xmath104 , and @xmath105 as defined before . we choose @xmath106 and @xmath107 in our main sets of simulations so that @xmath108 , similar to 2-d simulations above . @xmath109 is 1 and @xmath2 is 0.1 at @xmath110 . the grids are uniformly spaced in log(@xmath97 ) , @xmath99 , @xmath92 with 256@xmath67128@xmath67688 grid cells in the domain of [ log(0.3 ) , log(3)]@xmath67[@xmath111 - 0.6 , @xmath111 + 0.6 ] @xmath67[0 , 2@xmath112 for the main sets of simulations . in runs with @xmath62 and 100 , the cooling time decreases exponentially beyond @xmath113 with @xmath114 to mimic fast cooling at the disk surface ( dalessio et al . 1998 ) . numerically , this treatment also maintains better hydrostatic equilibrium at the disk surface . the boundary condition in the @xmath99 direction is chosen that @xmath115 in the ghost zones . we set @xmath116 and @xmath117 in the ghost zones having the same values as the last active zones . density in the ghost zones is set to be @xmath118 to maintain hydrostatic equilibrium in the @xmath99 direction , where @xmath119 and @xmath120 are the @xmath99 coordinates of the ghost and last active zones . we have also tried the boundary condition which sets the quantities in the ghost zones as the initial values , and found that the results are not affected by the choice of boundary conditions . to further study the numerical effect from the boundary , we have applied a wave damping zone ( de val - borro et al . 2006 ) operating at both @xmath97 and @xmath99 boundaries in the run sm2isolong . the wave damping zone in the @xmath97 direction is from @xmath121 to 1.25@xmath121 and from 0.84@xmath122 to @xmath122 . the damping zone in the @xmath99 direction is from @xmath123 to @xmath124 and from @xmath125 to @xmath126 . and @xmath121 , @xmath122 , @xmath123 , and @xmath126 are the boundary of the simulation domain . in these damping zones , the physical quantities are relaxed to the initial states on a timescale varying from infinity at the damping zone edge 1.25@xmath121 , 0.84@xmath122 , @xmath124 , and @xmath125 to a timescale of 0.1 orbit at the boundary of the simulation domain . the damping zone gradually damps waves traveling to the boundary of the simulation domain . besides the wave damping zone and the long timescale , the run sm2isolong is different from our main set of simulations in other ways . we ramp the planet mass linearly for 10 orbits in run sm2isolong to test how our results will be affected if we insert the planet gradually in the disk . we also choose @xmath127 in run sm2isolong to confirm that the small viscosity ( @xmath83 ) in the main set simulations will not affect the results . to verify our results , especially for the 3-d structure of spiral shocks , we have used athena with cylindrical coordinates ( run sm2isocyl ) to carry out a 3-d simulation having the same disk set - up as in sm2iso . this athena simulation ( sm2isocyl ) is different from the athena++ simulation ( sm2iso ) in several ways . first , sm2isocyl uses the corner transport upwind integrator which is different from the van - leer integrator used in athena++ . second , cylindrical coordinates with uniform radial grids have been used in sm2isocyl while spherical - polar coordinates with logarithmic radial grids have been used in sm2iso . the detailed disk set - up and the boundary conditions can be found in zhu et al . the results will be presented in 6.1 . overall , sm2isocyl confirms the results in sm2iso , which greatly limits the chance that our results are due to numerical artifacts . volume rendering of @xmath128 in simulation sm1iso is shown in figure [ fig : shockvender ] . @xmath129 is the density difference between 10 orbits and the initial condition , and @xmath130 is the initial density at that position . thus , @xmath128 highlights the density perturbation ( e.g. spiral shocks ) in the disk . in this paper , we use `` spiral shocks '' to refer to peaks of the density wakes and are associated with spiral arms seen in observations . it is apparent in figure [ fig : shockvender ] that the spiral shocks are not perpendicular to the disk midplane and they have complicated 3-d structure . in the figure , both the inner arms inside the planet and the outer arms outside the planet curl towards the central star at higher altitudes . this curled 3-d shock structure will be studied in more detail in 5 , while in this section we focus on the shape of the spiral wakes in the horizontal plane . figure [ fig : shock ] shows the shape of the spiral wakes in both 2-d and 3-d simulations . the x - axis is plotted in log @xmath64 , so that the pitch angle of the spiral wake can be easily estimated by using its slope in the figure ( d log @xmath64/d @xmath92 = tan @xmath19 ) . when a very low mass planet is present in a disk , it excites density waves that are in the linear regime . the linear theory for density waves in the 2-d @xmath131 plane ( equation [ eq : linearwake ] ) can accurately describe the shape of the excited spiral wakes in 2-d simulations . this is demonstrated in the upper left panel ( cm1iso ) of figure [ fig : shock ] where equation ( [ eq : linearwake ] ) with @xmath132 and @xmath133 fits the peak of the density wakes very well the excited density wakes become weak shocks at @xmath134 and 1.6 according to equation ( [ eq : eqshock ] ) . but the shocks are very weak and do not move away from the trajectory predicted by equation ( [ eq : linearwake ] ) significantly . ] . even at the midplane of 3-d simulations , equation ( [ eq : linearwake ] ) still provides a good fit to the density wakes ( the upper middle panel ) . however , the shape of the spiral arms at the disk surface is affected by the 3-d structure of density wakes . at the disk surface in 3-d simulations ( even in the linear regime , shown in the upper right panel of figure [ fig : shock ] ) , both inner and outer arms are at smaller @xmath64 than equation ( [ eq : linearwake ] ) due to the tilted shock shape in figure [ fig : shockvender ] . when these shocks are far away from the planet , they are more tilted towards the central star at higher altitudes . this leads to the inner spiral arms becoming slightly more open ( with a larger pitch angle ) and the outer spiral arms becoming slightly less open ( with a smaller pitch angle ) than equation ( [ eq : linearwake ] ) would predict . when the planet has a mass larger than @xmath32 ( middle and bottom panels of figure [ fig : shock ] ) , it can launch spiral shocks immediately around the planet , and the shape of spiral shocks can deviate from the trajectory predicted by equation ( [ eq : linearwake ] ) significantly . shocks excited by a more massive planet deviate from linear theory more and they have larger pitch angles . as shown in figure [ fig : shock ] ) , spiral shocks in the 6 @xmath11 cases ( bottom panels ) are more open and deviate from the prediction of linear theory ( dotted curves ) more than shocks in the 1 @xmath11 cases ( middle panels ) . the deviation from the linear theory has also been seen in previous simulations , e.g. , figure 2 and 10 of de val - borro et al . ( 2006 ) , but it has not been explored and the physical reason for the deviation is left to be unexplained . this deviation from linear theory shown in figure [ fig : shock ] is consistent with the predictions from the weakly non - linear density wave theory by goodman & rafikov ( 2001 ) and rafikov ( 2002 ) . in weakly non - linear theory , the spiral shock can expand in both azimuthal directions away from equation ( [ eq : linearwake ] ) , and , at each radius , the shock density profile along the azimuthal direction is n - shaped ( figure 2 in goodman & rafikov 2001 ) . the n - shaped shock profile expands in the azimuthal direction at a speed which is proportional to the normalized amplitude of the shock ( @xmath135 ) . thus , the higher is the planet mass , the stronger are the shocks and these shocks expand faster away from equation ( [ eq : linearwake ] ) . then , the spiral shock has a larger pitch angle as a result . similar to the 0.01 @xmath11 case , the inner spiral shocks in 1 and 6 @xmath11 cases are even more open at @xmath136 than at the midplane , while the outer arms become less open at the disk surface . for outer spiral arms , this 3-d effect compensates the increased pitch angel due to the shock expansion , and coincidently the outer arms almost overlap with the prediction from linear theory . another important feature shown in figure [ fig : shock ] is that , besides the primary inner arm which originates from the planet , a secondary inner spiral arm appears with some azimuthal shift from the primary arm . for some cases ( e.g. cm2iso ) , we even see a tertiary arm at the very inner disk . the secondary spiral arm has also been seen in previous simulations having massive planets , e.g. figure 2 in kley ( 1999 ) and figure 10 of de val - borro et al . however , it has hardly been explored in earlier simulations . figure [ fig : shock ] shows that , even with a very low mass planet ( 0.01 @xmath11 , the upper panels of figure [ fig : shock ] ) , another density peak ( the secondary arm ) emerges close to the primary inner arm with low density region ( the rarefaction wave of the primary arm ) in between . after the secondary inner arm appears close to the primary arm , it can become shock during the propagation and later it will become n - shaped which is similar to the primary arm . then in some cases , a tertiary inner arm appears at the rarefaction wave part of the secondary arm . for 0.01 @xmath11 cases , the primary and secondary inner arms are separated by @xmath137 at @xmath138 . when the planet gets more massive , the secondary inner arm is excited earlier and the separation between the primary and secondary arm increases . in 1 @xmath11 cases , the two inner arms are roughly separated by @xmath139 , and in 6 @xmath11 cases , the two arms are roughly separated by @xmath140 . this has an important application that we can use the separation between two arms to estimate the mass of the embedded planet . similar to the primary inner arm , the secondary inner arm is also stronger at the disk surface than at the disk midplane . for outer arms , the secondary arm appears in disks that have massive planets embedded , but the secondary outer arm is less apparent than the primary outer arm . more discussions on the 3-d structure of secondary arms will be presented in 5 . spiral wakes / shocks are slightly more open in a disk whose eos is not isothermal ( figure [ fig : shocktc1 ] ) . this is because density waves propagate slightly faster in a fluid with a non - isothermal eos than in a fluid at the same temperature with the isothermal eos ( e.g. goodman & rafikov 2001 ) . in figure [ fig : shocktc1 ] , even with a moderate cooling rate ( @xmath62 ) , the spiral wakes excited by a low mass planet ( 0.01 @xmath11 ) can only be fitted by equation ( [ eq : linearwake ] ) using a larger disk scale height that is calculated with the adiabatic sound speed instead of the isothermal sound speed ( @xmath141 in equation [ eq : linearwake ] is thus @xmath142 ) . it is a little bit surprising that the spiral shape in disks with @xmath62 follows the spiral shape in adiabatic disks . we think this is due to the short timescale for the flow to move across the spiral wake . the spiral wake has a typical width smaller than the disk scale height , while the background flow moves across the wake at nearly keplerian speed . thus , when the fluid travels in and out of the wave / shock , its response time is much smaller than the orbital time and thus behave adiabatically . other aspects of the spiral shocks in non - isothermal cases are similar to the isothermal cases , e.g. , a higher mass planet excites a more open spiral shock , and the inner spiral shocks become slightly more open at higher altitudes . to illustrate the shape of the spiral shocks in the physical space , we plot the relative density perturbation in cartesian coordinates in figure [ fig : shockcarg ] . clearly , the more massive the planet is , the more the spiral shocks deviate from linear theory . two well separated inner arms are also apparent when the planet mass is large , and the separation between these two arms is larger when the planet is more massive ( comparing the middle and bottom panels in figure [ fig : shockcarg ] ) . to see how successful the weakly non - linear density wave theory of goodman & rafikov ( 2001 ) and rafikov ( 2002 ) fits the shape of the shocks in simulations , we plot in figure [ fig : shockcut ] the density contour and density profiles along the azimuthal direction at @xmath138 , 0.5 , 1.5 and 2 for run cm2iso . in the density contour panel , we can clearly see that the secondary arm appears at the edge of the low density rarefaction wave region of the primary arm . after the secondary arm propagates for a short distance , a tertiary arm appears close to the low density rarefaction wave region of the secondary arm . in the panels showing the density profiles , we follow goodman & rafikov ( 2001 ) and rafikov ( 2002 ) to shift the density profiles so that @xmath143 corresponds to the wake position from linear theory ( black curve in figure [ fig : shockcut ] or equation [ eq : linearwake ] ) . the density profiles clearly show that the shocks are n - shaped . a rarefaction wave follows the shock front , and @xmath129 at the rarefaction wave region can be negative before it merges to the background flow . the shock fronts deviate from @xmath143 , and due to the shock expansion , the deviation is larger when the shock is further away from the planet . under the shearing - sheet approximation , the amplitude and width of the n - shaped shock scale as @xmath144 and @xmath145 at @xmath1460 based on the weakly non - linear density wave theory of goodman & rafikov ( 2001 ) . in a global disk spanning a large range of radii , the amplitude and width of the n - shaped shock scale as @xmath147 and @xmath148 , where @xmath149 is given in equation 43 of rafikov ( 2002 ) . for our disk parameters , we have @xmath150 since the middle point of the n - shaped shock is at @xmath1510 ( rafikov 2002 ) , we expect that in our global simulations the azimuthal distance between the shock front and the path predicted by equation [ eq : linearwake ] ( @xmath152=0 ) should also scales as @xmath148 . to test this prediction , we have measured the shock positions at @xmath153 and 1.5 in figure [ fig : shockcut ] , which are @xmath154 and 0.43 respectively . these two positions are labeled as the dashed lines in @xmath153 and 1.5 panels . then we calculate the shock positions at @xmath138 and 2 to be -1.7 and 0.9 , using their positions at @xmath153 and 1.5 together with the scaling relationship @xmath148 where @xmath149 at @xmath64=0.3 , 0.5 , 1.5 , and 2 are calculated from equation [ eq : tf ] . these predicted shock positions are labeled as the dashed lines in @xmath138 and 2 panels . we can see that they agree with the actual shock positions in the simulation very well . this confirms that the shock positions are determined by the non - linear expansion of spiral shocks . using the same approach , we have calculated the non - linear shock position at every radius for @xmath155 and @xmath156 with the normalization based on shock positions at @xmath153 and 1.5 . this new predicted shock shape is plotted as the dotted curve in the left panel of figure [ fig : shockcut ] . despite some offset at @xmath64 close to the planet , which is expected since the simple relationship @xmath147 holds only when @xmath1460 , a good agreement has been achieved between the non - linear density wave theory and the simulations . note that in this comparison , we did not calculate the shock strength directly from non - linear theory and compare its amplitude with simulations , instead we use the scaling relationship to verify the propagation of the shock . in future , direct comparison is desired when we have a more complete non - linear theory which can calculate the shock excitation directly . although the primary arm can be fitted by the weakly non - linear density wave theory , the excitation of the secondary ( or even tertiary ) arm still lacks a good theoretical explanation . it may be related to the low @xmath12 mode ( e.g. @xmath157 , @xmath158 , similar to disks in binary systems ) or some non - linear wave coupling . figure [ fig : shockcut ] suggests that a secondary spiral arm is excited at the other end of the n - shaped primary shock ( the @xmath153 panel ) . after it is excited , it steepens to shocks and becomes another n - shaped shock later ( the @xmath138 panel ) . its shock front can even travel into the rarefaction wave of the primary arm ( e.g. in the @xmath138 panel , the secondary shock is almost at @xmath152=0 where the rarefaction wave of the primary arm should reside . ) . unlike the primary arm which already dissipates when it travels inward from @xmath153 to 0.3 , this secondary arm is excited later and becomes stronger from @xmath153 to 0.3 . at @xmath138 , the secondary arm is even stronger than the primary arm . by comparing @xmath153 and @xmath138 panels , we also notice that the secondary arm almost keeps the same azimuthal separation with the primary arm ( @xmath1591.7 ) during its propagation . furthermore , at @xmath138 , a tertiary arm starts to appear at the other end of the n - shaped secondary arm . to summarize the results in this section , figure [ fig : pitch ] shows the pitch angle from the linear theory and those measured in numerical simulations . when the planet mass is low ( e.g. cm1iso , the green dots ) , the measured pitch angle agrees with the linear theory . when the planet mass increases , the pitch angle also increases . for the 6 @xmath11 case , the measured pitch angle of the spiral wake in the @xmath160 disk is close to the pitch angle predicted in a much thicker disk ( @xmath161 ) using the linear theory . the secondary and even the tertiary arms have similar pitch angles as the primary arms . thus , if we know the disk thermal structure very well , we can use the deviation of the measured pitch angle from the linear theory to estimate the embedded planet mass . since near - ir scattered light observations only probe the shape of the disk surface , 3-d structure of spiral shocks can affect the observational signatures of these spiral shocks . intuitively , we would expect that the spiral shocks have complicated 3-d structure . first , the wave excitation must have 3-d structure since , at the same @xmath64 in the disk , the distance between the planet and the disk surface is larger than the distance at the midplane , and the force is thus weaker at the disk surface . second , the wave propagation may have 3-d structure considering the disk becomes thinner at smaller @xmath64 . waves / shocks are more converged when they propagate inwards . they can also channel to the disk surface ( lubow & ogilvie 1998 , bate et al . 2002 , lee & gu 2015 ) , and , during their propagation from the high density region ( e.g. , the midplane ) to the low density region ( e.g. , the disk surface ) , the amplitudes of perturbations have to increase to conserve the wave action . since the amplitudes of perturbations determine when the waves will break into shocks , the dissipation can also be quite different between the surface and the midplane . all these effects can contribute to the 3-d structure of spiral waves / shocks . due to these complicated effects , it is difficult to develop an analytic theory to study the planet - induced 3-d shock structure , and we rely on numerical simulations to study such structure . however , before delving into the highly nonlinear shock regime , we can use the linear theory developed in tanaka , takeuchi & ward ( 2002 ) to estimate the 3-d effect of the density waves . following their theory for locally isothermal disks , the structure of the waves in the @xmath98 direction can be studied with hermite polynomials ( @xmath162 ) . we first expand perturbed quantities ( @xmath163 ) from our simulations into fourier series @xmath164\,,\ ] ] where the fourier components @xmath165 are complex functions of @xmath64 and @xmath98 . then , @xmath165 can be further expanded with hermite polynomials in the @xmath98 direction , @xmath166\,,\ ] ] where @xmath167 is the normalized height as @xmath168 , and the first three hermite polynomials are @xmath169 by using the normal orthogonal relation between @xmath170 , we have @xmath171 we can use @xmath172 at different @xmath173 to estimate the relative importance of different hermite components . to compare with figure 2 in tanaka et al . ( 2002 ) , we show @xmath174 , @xmath175 fourier - hermite components for the perturbed density ( @xmath128 ) in figure [ fig : m10 ] . with the same parameters , the right panel of figure [ fig : m10 ] is very similar to figure 2 in tanaka et al . ( 2012 ) . by comparing the left and right panels in figure [ fig : m10 ] , we can see that although higher - order hermite components have similar strength between thin and thick disks , n=0 hermite component is much weaker in a thick disk , suggesting that the wakes in a thick disk have more significant 3-d structures than wakes in a thin disk . figure [ fig : m10 ] suggests that higher - order vertical components can dominate the disk structure at the atmosphere . although it shows that the @xmath176 component ( @xmath177 ) is 10 times weaker than the @xmath178 component ( @xmath179 ) , and the @xmath178 component is 10 times weaker than the @xmath180 component ( which led tanaka et al . 2002 to conclude that most of the angular momentum excited by the planet will be carried by two dimensional free waves ) , the base function ( hermite polynomials ) at @xmath136 gets @xmath1 10 times larger sequentially from @xmath181 to @xmath182 and to @xmath183 . thus , @xmath184 and @xmath185 are still comparable with @xmath186 . the density structure at the disk atmosphere can be significantly affected by higher - order vertical modes . although the modal analysis is useful to verify numerical simulations and can be suggestive on the relative amplitudes of various modes , it is the 3-d structure in the real space that determines the observational signatures of waves / shocks . by studying the shock structure in real space , we first find that the 3-d shock structure is dramatically different between inner and outer arms . for the inner arms , the density perturbation is much larger at the disk surface than at the disk midplane . the 3-d structure of the inner spiral arms at @xmath153 is shown in figure [ fig : onedr0p5 ] . at @xmath153 and @xmath113 ( solid curves ) , the differences between the maximum and minimum density in the logarithmic scale are 0.015 , 0.4 , and 1.3 for m1 , m2 , and m3 cases respectively , in comparison with 0.004 , 0.2 , and 0.4 at the disk midplane ( dotted curves ) . at the same radius ( @xmath153 ) , the position of the wakes in non - isothermal disks ( lower panels ) are at smaller @xmath187 compared with those in isothermal disks ( upper panels ) . this is because the wakes are more open in non - isothermal disks , as discussed in 4 , . the secondary inner spiral arms / shocks are also more prominent at the disk surface than at the disk midplane . at the disk midplane , the secondary arms have lower amplitudes compared with the primary arms ( dotted curves in figure [ fig : onedr0p5 ] ) , while at @xmath113 , the secondary arms have almost the same amplitudes as the primary arms ( solid curves ) . the large amplitude of primary and secondary inner arms at the disk surface is due to the corrugated motion in the @xmath188 direction . in figure [ fig : onedr0p5 ] which is in the corotating frame with the planet , the disk material flows in the direction from the left side to the right side of the figure . before meeting with the shock , the disk is in vertical hydrostatic equilibrium with the background density and @xmath189 ( figure [ fig : vyr0p5 ] ) . after the shock , the disk material loses angular momentum and moves inwards with @xmath190 ( figure [ fig : vyr0p5 ] ) . at the same time , @xmath188 also becomes negative , compressing the disk material at the midplane . this downward motion decreases the density of the rarefaction wave at @xmath113 . before meeting the secondary shock , @xmath188 starts to increase and becomes positive , leading to a higher density at the disk surface . at the secondary shock , @xmath188 reaches the maximum positive velocity and leads to the highest density at the disk surface for the secondary shock in figure [ fig : onedr0p5 ] . this corrugated motion , first negative and then positive @xmath188 , leads to an enhanced contrast between the spiral shock and the rarefaction wave after the shock . on the other hand , the density perturbation of outer spiral arms is similar between the disk surface and the disk midplane , especially for isothermal disks , as shown in figure [ fig : onedr2 ] . at @xmath191 , regardless of height , the differences between the maximum and minimum density in the logarithmic scale are both 0.002 for sm1iso , 0.1 for sm2iso , and 0.3 for sm3iso . this lack of vertical variation is also reflected in figure [ fig : vyr2 ] where @xmath188 is very small compared with @xmath192 ( @xmath188 is almost two orders of magnitude smaller than @xmath192 ) . thus , the density structure of the outer spiral is mainly determined by @xmath192 and @xmath116 in the horizontal plane in stead of the corrugated motion in the @xmath99 direction . for non - isothermal runs ( bottom panels in figure [ fig : onedr2 ] ) , the density perturbation of the outer spiral arms at the disk surface is slightly higher than the perturbation at the disk midplane . disk material flows from the right hand side of the figure to the left hand side in figure [ fig : onedr2 ] and [ fig : vyr2 ] . when it meets the shock , it develops a @xmath188 towards the disk surface , which enhances the density at the disk surface . although it is tempting to contribute such difference between isothermal and non - isothermal runs to the nonlinear hydraulic jumps ( shock bores ) ( boley & durisen 2006 ) , such disk structure also appears even in the linear regime for the 0.01 @xmath11 case , implying that it may be a linear effect and related to the eigenfunctions of the 3-d waves excited by the planet . finally , to illustrate the increase of the density perturbation with height in disks and the qualitative difference between inner and outer arms , we plot the relative density perturbation along the radius at different heights ( @xmath1930 , 1 , 2 , 3 , 4 @xmath94 ) in figures [ fig : denmaxmin ] and [ fig : denmaxmintc1 ] . the relative density perturbation is defined as @xmath194 , where @xmath195 and @xmath196 are the maximum and minimum density along the azimuthal direction ( @xmath197 $ ] ) at the fixed @xmath64 and @xmath98 . we can see that the inner arms and outer arms are qualitatively different . for inner arms , the relative density perturbation is getting larger at higher altitudes , and the perturbation can increase by more than a factor of 10 from the midplane to the disk surface . for the outer arms , the relative density perturbation is almost unchanged between the midplane and the disk surface in isothermal disks ( figure [ fig : denmaxmin ] ) and only increases slightly from the midplane to the disk surface in non - isothermal disks ( figure [ fig : denmaxmintc1 ] ) . overall , for the inner arms , the large density perturbation at the disk surface has a significant effect on the near - ir observations as shown below . as shown in figure [ fig : shockvender ] , the 3-d spiral shocks are not perpendicular to the disk midplane . they curl in a way that they are almost along the @xmath99 direction in the spherical - polar grid . to quantify the curl , the upper panel of figure [ fig : oned ] shows the relative density perturbation at the disk midplane ( black curves ) and at @xmath198 ( cyan curves ) for run sm2iso ( solid curves ) . please note that the x - axis in figure [ fig : oned ] is @xmath97 ( the radial position in spherical - polar coordinates ) instead of @xmath64 ( the radial position in cylindrical coordinates ) . the fact that the shock density peaks at almost the same @xmath97 at either the disk midplane or @xmath198 demonstrates that the shocks almost curl along the @xmath99 direction . such coincidence makes us suspect that they could be numerical artifacts due to the adopted spherical - polar grid structure . thus , we use athena to carry out a similar simulation but using cylindrical coordinates ( run sm2isocyl introduced in 3.3 ) . after transforming simulation outputs from cylindrical coordinates to spherical - polar coordinates , the relative density perturbation for sm2isocyl is shown in the upper panel of figure [ fig : oned ] as the dotted curves . the shocks are at similar positions as those in sm2iso , which confirms that the curled shock structure is real instead of numerical artifacts . the physical reason for the curled shock structure is unclear . one would suspect that it may be related to the radial temperature gradient in the local isothermal disk . in such disks , the disk rotates at different angular velocities at different disk heights . such vertical shear can change disk dynamics , such as leading to the vertical shear instability ( nelson et al . 2013 ) and it may also lead to curled shock structures . to test this idea , we have run one simulation with a constant temperature in the whole disk ( sm2isoconstt ) . in such a disk , the disk rotates at a constant angular velocity at a given @xmath64 independent on the disk height . the relative density perturbation is shown in the bottom panel of figure [ fig : oned ] . however , the shock density still peaks at almost the same @xmath97 at different @xmath99 , implying that the curl of the shock is not caused by the vertical shear in disks . thus , the physical mechanism for the curled shock remains unclear and still deserves further explore . to demonstrate that our derived shock structure is independent on specific numerical choices in our model , we show the density perturbation for run sm2isolong in figure [ fig : longterm ] . run sm2isolong is different from sm2iso in several aspects ( as introduced in 3.3 ) : 1 ) the simulation runs for 120 orbits instead of 10 orbits ; 2 ) it has a wave damping region , 3 ) the planet mass is ramped up slowly over 10 orbits , and 4 ) it uses zero viscosity . despite these differences , its density perturbation at 15 orbits ( bottom panel of figure [ fig : longterm ] ) is quite similar to the density perturbation in sm2iso ( middle panel of figure [ fig : denmaxmin ] ) . when a gap is induced at 120 orbits , the spiral shocks become weaker at the disk region close to the gap , while it still maintains the same strength at the region far away from the gap . we notice that vortices start to develop at the outer gap edge ( @xmath199 ) at 120 orbits , which slightly enhances the density perturbation at the outer gap edge . we have also studied how the shock structure can be affected by different disk structures . figure [ fig : thindisk ] shows the relative density perturbation of the wakes in a thinner disk ( sthin ) . compared with figure [ fig : denmaxmin ] , we can see that the density perturbation increases by a factor of 2 from the midplane to 3 @xmath94 , and a factor of 4 from the midplane to 4 @xmath94 in the thin disk , compared with a factor of 4 and 10 respectively in the thick disk . this suggests that the wakes / shocks have more significant 3-d structure in a thicker disk . although this does not favor detecting spiral arms in thinner disks in future , we need to keep in mind that the thermal mass ( equation [ eq : thermal ] ) is smaller in a thinner disk so that the same mass planet corresponds to a more massive planet in the scale of thermal mass and it will excite stronger density waves . in this case the inner arms may still be observable in a thin disk . to understand how the 3-d structure of density waves / shocks affects observations , we post - process our hydrodynamical simulations with monte - carlo radiative transfer calculations to generate near - ir scattered light images ( figures [ fig : image6mjfull ] , [ fig : image1mj ] and [ fig : image6mj ] ) . the details on the monte - carlo radiative transfer calculations are presented in dong et al . ( 2014 , 2015 ) . to assign physical scales to our simulations , we assume that the planet is at 50 au and the central source is a typical herbig ae / be star ( 2 m@xmath200 ) with a temperature of 10@xmath201 k and a radius of 2@xmath202 . ism dust grains have been used and their distribution is assumed to follow the gas distribution . following dong et al . ( 2015 ) , we assume that the gas disk is 0.02 @xmath7 . with the gas to dust mass ratio of 100:1 , the total dust mass is 2@xmath203 . further assuming 10% of dust is in the form of ism dust , the total mass of the ism dust is 2@xmath204 . in this model , toomre q parameter at 50 au is @xmath130 , so neglecting disk self - gravity is justified . in mcrt simulations , photons from the central star are absorbed / reemitted or scattered by the dust in the surrounding disk . the convolved images in figures [ fig : image6mjfull ] , [ fig : image1mj ] and [ fig : image6mj ] are derived by convolving full resolution images with a gaussian point spread function having a full width half maximum ( fwhm ) of 0.06 . this resolution is comparable with nir direct imaging observations using subaru , vlt , and gemini . in the right two panels of figures [ fig : image1mj ] and [ fig : image6mj ] , we assume that the object is 140 pc away , while in figure [ fig : image6mjfull ] and the left two panels of figures [ fig : image1mj ] and [ fig : image6mj ] we assume that the distance is 70 pc so that the inner arms are shown more clearly . both full intensity and polarized intensity images are calculated and shown in figure [ fig : image6mjfull ] . the spiral arms are evident in these images . when the disk is viewed face - on , both full intensity and polarized intensity images are almost identical except that the full intensity image is almost a factor of 2 brighter than the polarized intensity image . the polarization fraction , which is defined as the ratio between the polarized intensity and full intensity , is almost a constant ( @xmath10.45 ) . when the disk is viewed at some inclination angle , the polarization fraction is not a constant due to the dust forward scattering . the dependence of the polarization fraction on disk inclination can be useful to determine the disk inclination angle . to highlight the importance of the 3-d wave structure , we have also computed models by only using the disk midplane density from simulations , which is labeled as 2d@xmath2053d in figures [ fig : image1mj ] and [ fig : image6mj ] . in these models , we assume that the disk is in vertical hydrostatic equilibrium and puff up the midplane density to higher altitudes as @xmath206 where @xmath207 ( the same as the scale height used in 3-d simulations ) . figures [ fig : image1mj ] and [ fig : image6mj ] show that , in 3d models , the inner arms are considerably more prominent than the outer arms , and normally a secondary arm can be as bright as the primary arm . the shape of the inner arms clearly deviate from the prediction of linear theory . as expected from the non - linear expansion of spiral shocks , the pitch angle of the inner spiral arms in the more massive planet case ( sm3iso , figure [ fig : image6mj ] ) is larger than those in the less massive planet case ( sm2iso , figure [ fig : image1mj ] ) . the inner arms are also quite sharp , while the outer arms are quite broad and sometimes indistinguishable from the background disk . this difference is partly because the sharp shock fronts are facing the star for the inner arms , while they are facing away from the star for the outer arms . the different geometry at the disk surface can greatly affect the intensity of the scattered light images ( takami et al . we calculate an approximate scattering surface , defined as the disk surface where the column density is 0.01 ( in code units ) , for the sm3iso model at @xmath208 , shown in figure [ fig : scattering ] . clearly , for the inner arms , the shock fronts are facing the star , while , for the outer arms , the smooth rarefaction waves are facing the star . since the rarefaction waves change gradually with radius , they are illuminated by the star more uniformly than the shock . thus , the outer arms appear quite broad . however , when the planet mass is not very high ( 1 m@xmath209 case ) , the width of rarefaction waves in the radial direction can be smaller than the size of the observational beam , and we wo nt be able to distinguish the inner and outer arms based on the sharpness of the arms . in 3-d models of figure [ fig : image1mj ] and [ fig : image6mj ] , the secondary inner arms are as apparent as the primary arms , even though the primary arms have higher surface density than secondary arms . this is due to the corrugated motion discussed above in figure [ fig : onedr0p5 ] and [ fig : vyr0p5 ] , which increases the density of the secondary arms at the disk surface . the secondary arm is offset from the primary arm with some azimuthal angle , as also shown in the surface density plot ( figure [ fig : shock ] ) . this offset is smaller in sm1iso ( figure [ fig : image1mj ] ) than that in sm6iso ( figure [ fig : image6mj ] ) . the two spiral arms are @xmath210 apart in sm1iso ( figure [ fig : image1mj ] ) , while almost 180@xmath211 apart in the more massive planet case ( sm6iso , figure [ fig : image6mj ] ) . by comparing 3d models with 2d@xmath2053d models in figure [ fig : image1mj ] and [ fig : image6mj ] , we find that the inner spiral arms are more prominent in 3d models , as expected due to inner arms higher density perturbation at the disk surface in 3d models ( figure [ fig : denmaxmin ] ) . even in convolved images , the polarized intensity of inner arms is at least twice stronger in 3d models than in 2d@xmath2053d models . the outer shocks are not very apparent , and they are similar between 3d and 2d@xmath2053d models since the density perturbation of outer arms is almost height independent ( 5 ) . as discussed in 4 , the outer arms coincidently follow linear theory ( equation [ eq : linearwake ] ) . when the planet mass is very large as in figure [ fig : image6mj ] , the secondary outer arm starts to become visible . another noticeable difference between 3d and 2d@xmath2053d models is that the planet ( or the circumplanetary region ) is bright in 2d@xmath2053d models while it is dim in 3d models . this is because , when we puff the disk from 2-d to 3-d , we have ignored the planet s gravity so that the higher density in the circumplanetary region leads to the higher density at the disk surface . the circumplanetary region even casts a shadow to the outer disk in 2d@xmath2053d models . in realistic 3-d models , the gravity of the planet has been self - consistently included , which pulls the circumplanetary material towards the disk midplane and leads to a lower density at the disk surface . thus , the circumplanetary region receives less irradiation by the central star , and becomes dark . on the other hand , the outer disk beyond the planet is better illuminated and thus becomes bright instead of being shadowed . however , we need to keep in mind that we have ignored the luminosity from the planet and the circumplanetary disk in our models . zhu ( 2015 ) point out that accreting circumplanetary disks can be very bright ( @xmath10.001 @xmath8 if the circumplanetary disk accepts onto jupiter at a rate @xmath212 ) . such high luminosity may be able to illuminate the circumplanetary region significantly . we may also be able to directly detect such accreting circumplanetary disks in direct imaging observations operating at mid - ir wavelengths . previous works suggest that spiral patterns from direct imaging observations can not be explained by planet - induced spiral arms , since the pitch angle of the spiral arm is larger in observations than that predicted by the linear spiral wave theory , and the contrast of the spiral arm is higher in observations than suggested by the synthetic observations based on two dimensional planet - disk simulations . however , most previous works only focus on using outer spiral arms beyond the planet to explain observed spiral patterns . as shown in figure [ fig : pitch ] , the inner arms generally have larger pitch angle than outer arms even under the linear density wave theory . furthermore , when the spiral arms become spiral shocks , the pitch angle increases with stronger shocks . we also found that a secondary ( or even a tertiary ) spiral arm , especially for the inner arms , is also excited by a massive planet . the more massive is the planet , the larger is the separation in the azimuthal direction between the primary and secondary arms . the inner arms also have significant vertical motion , which boosts the density perturbation at the disk surface and the intensity contrast in synthetic images . thus , we have three independent ways to estimate the planet mass based on the spiral patterns in observations : 1 ) using the deviation between the measured pitch angle and the pitch angle predicted in the linear theory ( figure [ fig : pitch ] ) ; 2 ) the existence of the secondary spiral arm and the separation between two arms ( figure [ fig : shock ] ) ; 3 ) the intensity contrast of the spiral arms in observation ( figure [ fig : image1mj ] & [ fig : image6mj ] ) . the first two methods only use the information on the shape of the spiral arms and are less likely to be affected by the disk vertical temperature structure . in a companion paper ( dong et al . 2015 ) , we have shown that all these three methods suggest that sao 206462 and mwc 758 may harbor massive planets ( several to tens of jupiter masses ) outside the detected spiral arms . the planet - induced spiral arms can efficiently scatter light from the central star and are quite bright within gaps ( figure [ fig : image6mjfull ] ) . the spiral arms may be responsible for some resolved infrared features within the cavity of lkca 15 ( kraus & ireland 2012 ) . we have carried out two dimensional ( 2-d ) and three dimensional ( 3-d ) hydrodynamical simulations to study spiral wakes / shocks excited by young planets . simulations with different planet masses ( 0.01 , 1 , and 6 @xmath11 ) and different equations of state ( isothermal and adiabatic ) have been carried out . * we find that the linear density wave theory can only explain the shape of the spiral wakes excited by a very low mass planet ( e.g. 0.01 @xmath11 ) . spiral shocks excited by high mass planets clearly deviate from the prediction of linear theory . for a more massive planet , the deviation is more significant and the pitch angle of the spiral arms becomes larger . this phenomenon can be nicely explained by the wake broadening from the non - linear density wave theory ( goodman & rafikov 2001 , rafikov 2002 ) . a more massive planet excites a stronger shock which expands more quickly , leading to a larger pitch angle . * a secondary ( or even tertiary ) inner spiral arm is also excited by the planet . it seems to be excited at the edge of the n - shaped primary arm . the more massive is the planet , the larger is the separation between the primary and secondary arm . at the disk surface , the secondary inner arm can be as strong as the primary arm . the secondary inner arm almost keeps the same azimuthal separation with the primary arm at every radius in the disk . the excitation mechanism for the secondary ( tertiary ) arm deserves further study . * the spiral shocks have significant 3-d structure . they are not perpendicular to the disk midplane . they are curled towards the star at the disk surface . this further increases the pitch angle of the inner arms at the disk surface , but reduces the pitch angle of the outer arms at the disk surface . for outer arms , this effect compensates the increased pitch angle due to wake broadening . eventually at the disk surface , the shape of outer spiral arms still roughly follows the prediction of linear theory , while the inner arms are considerably more opened than predicted by the linear theory . * the inner spiral shocks also have significant vertical motion . the corrugated motion increases the density perturbation of the inner spiral arms by more than a factor of 10 at @xmath213 compared with the perturbation at the disk midplane . this can dramatically increase the contrast of the spiral patterns in near - ir scattered light images . the outer spiral shocks have little vertical motion in isothermal disks . with a non - isothermal eos , there are some vertical motions for the outer arms , which can make the outer arms more apparent . * we have combined our hydrodynamical simulations with monte - carlo radiative transfer calculations to generate near - ir scattered light images . we find that the inner spiral arms are prominent features that are observable by current near - ir imaging facilities . besides the apparent spiral patterns in some transitional disks , planet - induced spiral arms may also be responsible to some marginally detected near - ir features within cavity of some transitional disks . we further demonstrate that inner spiral arms in synthetic near - ir images using full 3-d hydrodynamical models are much more prominent than those based on 2-d models assuming hydrostatic equilibrium , consistent with the 3-d structure of the inner arms . on the other hand , the outer shocks are not very apparent and they are similar in synthetic images using 3d and 2d models . this indicates the need to model observations ( especially for inner arms ) with full 3-d hydrodynamics . * the different geometry between the inner and outer arms also affects their appearance in near - ir images . the sharp shock fronts of the inner arms face the central star directly , producing sharp narrow spiral features in observations . on the other hand , for the outer arms , the smooth rarefaction waves face the central star , producing broad and dimmer spiral features . * in near - ir images , the circumplanetary region is very dim since the planetary gravity reduces the density at the disk atmosphere . however , the disk region behind the planet can be better illuminated and becomes bright . in the appendix , we have shown that buoyancy resonances are confirmed in global adiabatic simulations even if the disk has a moderate cooling rate . they can lead to sharp density ridges around the planet , which may have observational signatures . overall , spiral arms ( especially inner arms ) excited by low mass companions are prominent features in near - ir scattered light images . most importantly , we have three independent ways to infer the companion s mass : 1 ) the pitch angle of the spiral patterns , 2 ) the separation between the primary and secondary arms , and 3 ) the contrast of the spiral patterns . in a companion paper ( dong et al . 2015 ) , we have combined mcrt and hydrodynamical simulations from this paper , and shown that planet - induced inner arms can explain spiral patterns revealed by recent near - ir direct imaging observations for sao 206462 and mwc 758 . we want to caution that our proposed model can not explain the gaps at the inner disks . other mechanisms ( e.g. another planet or disk photoevaporation ) have to be invoked to explain the inner gaps . all hydrodynamical simulations are carried out using computer supported by the princeton institute of computational science and engineering , and the texas advanced computing center ( tacc ) at the university of texas at austin through xsede grant tg- ast130002 . this project is supported by nasa through hubble fellowship grants hst - hf-51333.01-a ( z.z . ) and hst - hf-51320.01-a ( r.d . ) 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 . the inner and outer spiral shocks are due to the steepening of spiral density waves which are excited by the planet at lindblad resonances . at lindblad resonances , the doppler shifted frequency matches the disk epicyclic frequency ( @xmath214 ) and density waves are excited . however , besides the epicyclic frequency , the disk also has other natural frequencies . when the disk is not strictly isothermal , it has a non - zero brunt - visl frequency @xmath215 where @xmath216 . matching the brunt - visl frequency with the doppler shifted frequency , we have @xmath217 given a @xmath12 , equation ( [ eq : buoy ] ) gives the position of the resonances . these buoyancy resonances were discovered in shearing box simulations ( zhu , stone , & rafikov 2012 ) and studied analytically in lubow & zhu ( 2014 ) . they have significant contributions to the planetary torque , especially around the planet , which may affect planet migration . these resonances are infinitely thin , and no waves are excited to carry the deposited angular momentum and energy away . their dissipation relies on microscopic viscosity or radiative cooling . thin density ridges with large temperature and velocity variations appear at these resonances . when various @xmath12 modes overlap with each other , we can roughly estimate the position of the final density ridges caused by buoyancy resonances following equation 10 and 11 in zhu et al . first , given a @xmath12 , we can calculate the corresponding resonance position at @xmath64 and @xmath98 . then using the azimuthal wavelength for this mode @xmath218 , the geometric location of the constant phase @xmath219 ( @xmath173 is integer ) is given by @xmath220 ( assuming the phase of buoyancy waves is 0 at the planet position . ) . thus , @xmath221 we plot temperature fluctuations and @xmath188 at @xmath222 for sm1t1 in figure [ fig : buoy ] . the positions of buoyancy resonances given by equation ( [ eq : buoyt ] ) are plotted with dotted lines / curves . figure [ fig : buoy ] shows that both temperature fluctuations and @xmath188 are nicely tracked by equation ( [ eq : buoyt ] ) , suggesting that buoyancy resonances exist in disks even with @xmath62 . even though buoyancy resonances can affect planet migration , they may not be observed through direct imaging technique since the density fluctuations caused by these resonances are much weaker than the spiral density waves excited by lindblad resonances . for example , in the @xmath136 panels of figure [ fig : shocktc1 ] , we can see some density fluctuations close to the corotation region ( especially for sm2t1 , the 1 @xmath11 case ) , but they are much weaker than the spiral shocks . these buoyancy resonances also have vertical structure , as shown in figure [ fig : temp2d ] . we sliced through the disk at a fixed @xmath92 , and the buoyancy resonance curves are plotted as dotted curves in figure [ fig : temp2d ] . at the disk midplane , @xmath223 and there are no buoyancy resonances . since @xmath224 increases with disk height , @xmath225 also needs to increase with height to match the brunt - visl frequency considering @xmath226 and m are the same at the same @xmath92 . thus , the curves move away from the planet position towards the disk atmosphere .	direct imaging observations have revealed spiral structures in protoplanetary disks . previous studies have suggested that planet - induced spiral arms can not explain some of these spiral patterns , due to the large pitch angle and high contrast of the spiral arms in observations . we have carried out three dimensional ( 3-d ) hydrodynamical simulations to study spiral wakes / shocks excited by young planets . we find that , in contrast with linear theory , the pitch angle of spiral arms does depend on the planet mass , which can be explained by the non - linear density wave theory . a secondary ( or even a tertiary ) spiral arm , especially for inner arms , is also excited by a massive planet . with a more massive planet in the disk , the excited spiral arms have larger pitch angle and the separation between the primary and secondary arms in the azimuthal direction is also larger . we also find that although the arms in the outer disk do not exhibit much vertical motion , the inner arms have significant vertical motion , which boosts the density perturbation at the disk atmosphere . combining hydrodynamical models with monte - carlo radiative transfer calculations , we find that the inner spiral arms are considerably more prominent in synthetic near - ir images using full 3-d hydrodynamical models than images based on 2-d models assuming vertical hydrostatic equilibrium , indicating the need to model observations with full 3-d hydrodynamics . overall , companion - induced spiral arms not only pinpoint the companion s position but also provide three independent ways ( pitch angle , separation between two arms , and contrast of arms ) to constrain the companion s mass .
first introduced by gallager @xcite , ldpc codes have been the focus of intense research in the past decade and many of their properties are now well - understood . the iterative decoding algorithms for ldpc codes have been analyzed in detail , and asymptotic performance results have been derived @xcite . however , estimation of frame - error - rate ( fer ) for iterative decoding of finite - length ldpc codes is still an unsolved problem . a special case of interest is the performance of iterative decoding at high signal - to - noise ratio ( snr ) . at high snrs , a sudden degradation in the performance of iterative decoders has been observed @xcite,@xcite . this abrupt change manifested in the fer curve is termed as an `` error - floor . '' the error - floor problem is well - understood for iterative decoding over the binary erasure channel ( bec ) @xcite . combinatorial structures called `` stopping sets '' were used to characterize the fer for iterative decoding of ldpc codes over the bec . it was established that decoding failure occurs whenever all the variables belonging to stopping sets are erased . @xcite used this fact to construct irregular ldpc codes which avoid small stopping sets thus improving the guaranteed erasure recovery capability of codes under iterative decoding , and hence improving the error - floors . as in the case of bec , a strong connection has been found between the existence of low - weight uncorrectable error patterns and error - floors for additive white gaussian noise ( awgn ) channels and binary symmetric channels ( bsc ) ( see @xcite and @xcite ) . hence , studying the guaranteed error correction capability of codes under iterative decoding is important in the context of characterization and improvement of the performance of iterative decoding strategies . in the past , guaranteed error correction has been approached from the perspective of the decoding algorithm as well as from the perspective of code construction . sipser and spielman @xcite used expansion arguments to derive sufficient conditions for the parallel bit - flipping algorithm to correct a fraction of errors in codes with column - weight greater than four . burshtein @xcite proved that for large enough lengths , almost all codes with column - weights greater than or equal to four can correct a certain fraction of errors under the bit - flipping algorithm . burshtein and miller @xcite derived the sufficient conditions for message - passing decoding to correct a fraction of errors for codes of column - weight greater than five . however , these proofs were not constructive , i.e. , no explicit code construction which satisfied the sufficient conditions was provided . moreover , the code - lengths required to guarantee the correction of a small number of errors ( say @xmath0 ) is very high . also , these arguments can not be extended for message - passing decoding of codes with column - weight three or four . in order to construct codes with good error correcting properties under iterative decoding , progressive edge growth ( peg ) @xcite and constructions based on finite geometries @xcite have been used . however , codes constructed from finite geometries typically have very high column - weight . although , it has been proved that minimum distance grows at least linearly for codes constructed using peg , no results proving guaranteed error correction under iterative decoding exist for these codes . in this work , we derive necessary and sufficient conditions for the correction of three errors in a column - weight - three code under the hard - decision message - passing algorithm . we provide a modified peg construction which yields codes with such an error - correction capability . also , we derive the necessary and sufficient conditions for the correction of three errors in four iterations for the case of codes with column - weight four . again , we provide a modified peg construction which yields codes with such error - correction capability . the remainder of the paper is organized as follows : we establish the preliminaries of the work in section [ section2 ] . the necessary and sufficient conditions for the correction of three errors in column - weight - three codes are derived in section [ section3 ] . the case of column - weight - four codes is dealt with in section [ section4 ] . in section [ section5 ] , we describe a technique to construct codes satisfying the conditions of the theorems and provide numerical results . we conclude with a few remarks in section [ section6 ] . in this section , we first describe the tanner graph representation of ldpc codes . then , we establish the notation that will be used throughout this paper . finally , we describe the hard - decision message - passing algorithm that will be used for decoding . the tanner graph of an ldpc code , @xmath1 , is a bipartite graph with two sets of nodes : @xmath2 , the variable ( bit ) nodes and @xmath3 , the check ( constraint ) nodes . every edge @xmath4 in the bipartite graph is associated with a variable node @xmath5 and a check node @xmath6 . the check nodes ( variable nodes , respectively ) connected to a variable node ( check node , respectively ) are referred to as its neighbors . the degree of a node is the number of its neighbors . in a @xmath7-regular ldpc code , each variable node has degree @xmath8 and each check node has degree @xmath9 . the girth @xmath10 is the length of the shortest cycle in @xmath11 . let @xmath12 such that @xmath13 . if for all choices of @xmath14 , there are at least @xmath15 neighbors of @xmath14 in @xmath3 , then we say that the @xmath16 condition is satisfied . in this paper , @xmath17 represents a variable node , @xmath18 represents an even - degree check node and @xmath19 represents an odd - degree check node . let @xmath20 $ ] , a binary @xmath21-tuple , be the input to the message - passing decoder . let @xmath22 be a variable node with @xmath23 as its corresponding bit and @xmath24 be a check node neighboring @xmath5 . let vcj denote the message that @xmath5 sends to @xmath6 in the first half of the @xmath25 iteration and cvj denote the message that @xmath6 sends to @xmath5 in the second half of the @xmath25 iteration additionally , let v : j be the set of all messages from a variable @xmath5 to all its neighboring checks in the first half of the @xmath25 iteration . let v:\cj be the set of all messages that a variable node @xmath5 sends to all its neighboring checks except @xmath6 in the first half of the @xmath25 iteration . let @xmath26 be the set of all messages received by @xmath5 from all its neighboring in the second half of the @xmath25 iteration . let @xmath27 be the set of all messages received by @xmath5 from all its neighboring check nodes except @xmath6 in the second half of the @xmath25 iteration . c : j , c:\vj , @xmath28 and @xmath29 are defined similarly . the gallager algorithms @xcite can be defined as follows : the forward messages , vcj ( from variables to checks ) , are defined as @xmath30 where @xmath31 refers to the total number of messages which are of the value @xmath32 . the backward messages , cvj ( from checks to variables ) , are defined as @xmath33 at the end of each iteration , an estimate of each variable node is made based on the incoming messages and possibly the received value . the decoder is run until a valid codeword is found or until a maximum number of iterations , say @xmath34 , is reached , whichever is earlier . in eqn . ( [ equation : forward ] ) , @xmath35 is a threshold which is generally a function of the iteration number , @xmath36 , and the degree of the variable @xmath5 . in this paper , we use @xmath37 for all @xmath5 and @xmath36 for decoding column - weight - three codes . for column - weight - four codes , we use @xmath38 for all @xmath5 when @xmath39 and @xmath37 for all @xmath5 when @xmath40 . _ remark : _ we note that eqns . [ equation : forward ] and [ equation : backward ] then correspond to the gallager - b algorithm @xcite . for the gallager - a algorithm @xcite , @xmath41 , for all @xmath36 , where @xmath42 is degree of variable node @xmath5 . + _ a note on the decision rule : _ different rules to estimate a variable node after each iteration are available , and it is likely that changing the rule after certain number of iterations may be beneficial . however , the analysis of such scenarios is beyond the scope of this paper . throughout the paper , we use the following decision rule : if all incoming messages to a variable node from neighboring checks are equal , set the variable node to that value ; else set it to its received value . we discuss briefly the concept of trapping sets . consider an ldpc code of length @xmath21 . let @xmath43 be the binary vector which is the input to the hard - decision decoder . for output symmetric channels , without loss of generality , we can assume that the all - zero - codeword is transmitted . we make this assumption throughout this paper . the support of a vector @xmath43 denoted by @xmath44 is defined as the set of all positions @xmath45 where @xmath46 . for each @xmath47 , @xmath48 , let @xmath49 be the codeword estimate of the decoder at the end of the @xmath50 iteration . a variable node @xmath5 is said to be _ eventually correct _ if there exists a positive integer @xmath51 such that for all @xmath52 , @xmath5 does not belong to @xmath53 . @xcite a decoding failure is said to have occurred if there does not exist @xmath54 such that @xmath55 @xcite let @xmath56 denote the set of variable nodes that are not eventually correct . if @xmath56 is not empty , let @xmath57 and @xmath58 be the number of odd - degree check nodes in the subgraph induced by @xmath56 . we say that @xmath56 is an @xmath59 trapping set . let @xmath60 be a trapping set and let @xmath61 . the critical number @xmath62 of trapping set @xmath60 is the minimum number of variable nodes that have to be initially in error for the decoder to end up in the trapping set @xmath60 . that is , @xmath63 . @xcite let @xmath60 be a trapping set . if @xmath64 , then @xmath44 is a failure set of @xmath60 . for transmission over the bsc , @xmath43 is a fixed point of the decoding algorithm if @xmath65 for all @xmath47 . it follows that for transmission over the bsc , if @xmath43 is a fixed point , then @xmath66 is a trapping set . now , we have the following theorem which provides the sufficient condition for a set of variables to be a trapping set : [ thm1]@xcite let @xmath1 be the tanner graph of a column - weight - three code . let @xmath67 , be a set consisting of @xmath5 variable nodes with induced subgraph @xmath68 . let the checks in @xmath68 be partitioned into two disjoint subsets , namely , @xmath69 consisting of checks with odd degree and @xmath70 consisting of checks with even degree . if ( a ) every variable node in @xmath68 is connected to at least two checks in @xmath70 and at most one check in @xmath69 and ( b ) no two checks of @xmath69 are connected to the same variable node outside @xmath68 , then @xmath60 is a trapping set . see @xcite . in this section , we establish necessary and sufficient conditions for a column - weight - three code to correct three errors . we first illustrate three trapping sets and show that the critical number of these trapping sets is three thereby providing necessary conditions to correct three errors . we then prove that avoiding structures isomorphic to these trapping sets in the tanner graph is sufficient to guarantee correction of three errors . [ trappingsets ] shows three subgraphs induced by different numbers of variable nodes . let us assume that in all these induced graphs , no two odd degree checks are connected to the same variable node outside the graph . by the conditions of theorem [ thm1 ] , all these induced subgraphs are trapping sets . fig . [ sixcycle ] is a @xmath71 trapping set , fig . [ 53trappingset ] is a @xmath72 trapping set and fig . [ weight8codeword ] is a @xmath73 trapping set . note that a @xmath71 trapping set is isomorphic to a six - cycle and the @xmath73 trapping set is a codeword of weight eight . we now have the following result : [ sixcycle ] [ 53trappingset ] [ weight8codeword ] the critical number for a @xmath71 trapping set which is also a fixed point is at most three . there exist ( 5,3 ) and ( 8,0 ) trapping sets with critical number three and no ( 5,3 ) or ( 8,0 ) trapping sets with critical number less than three . [ vc1 ] [ cv1 ] [ vc2 ] [ cv2 ] the proof for @xmath71 case is trivial . we prove the lemma for the case of @xmath72 trapping sets and omit the proof for @xmath73 trapping sets . consider the @xmath74 trapping set shown in fig . [ messagepassing ] . let @xmath75 be the set of variables which are initially in error . let @xmath76 and @xmath77 . also , assume that no variable node in @xmath78 , has two or more neighbors in @xmath79 . in the first iteration , we have : @xmath80 consequently , all variable nodes in @xmath81 are decoded incorrectly at the end of the first iteration . in the second iteration : @xmath82 and all variable nodes in @xmath83 are decoded incorrectly . continuing in this fashion , @xmath84 and @xmath85 that is , the messages being passed in the tanner graph would repeat after every two iterations . hence , three variable nodes in error initially can lead to a decoder failure and therefore , this @xmath74 trapping set has critical number equal to three . to guarantee that three errors in a column - weight - three ldpc code can be corrected by the gallager - a algorithm , it is necessary to avoid @xmath71 , @xmath72 and @xmath73 trapping sets in its tanner graph . follows from the discussion above . we now state and prove the main theorem . if the tanner graph of a column - weight - three ldpc codes has girth eight and does not contain a subgraph isomorphic to a @xmath72 trapping set or a subgraph isomorphic to an @xmath73 trapping set , then any three errors can be corrected using the gallager - a algorithm . let @xmath86 be the three erroneous variables and @xmath79 be the set of the checks connected to the variables in @xmath83 . in a column - weight - three code ( free of cycles of length four ) the variables in @xmath83 can induce only one of the five subgraphs given in fig . [ errorconfigs ] . in each case , @xmath87 if @xmath88 and is @xmath89 otherwise . the proof proceeds by examining these subgraphs one at a time and proving the correction of the three erroneous variables in each case . [ config1 ] [ config2 ] [ config3 ] [ config4 ] [ config5 ] * subgraph 1 : * since the girth of the code is eight , it has no six cycles . hence , the configuration in fig . [ config1 ] is not possible . * subgraph 2 : * the variables in @xmath83 induce the subgraph shown in fig . [ config2 ] . at the end of the first iteration : @xmath90 there can not exist a variable node which is connected to two or more checks in the set @xmath79 without introducing either a six - cycle or a subgraph isomorphic to @xmath72 trapping set . at the end of first iteration , @xmath91 for all @xmath88 . furthermore , there exists no @xmath92 for which @xmath93 . hence , if a decision is made after the first iteration , a valid codeword is found and the decoder is successful . * subgraph 3 : * the variables in @xmath83 induce the subgraph shown in fig . [ config3 ] . at the end of the first iteration : @xmath94 for no @xmath92 , @xmath93 as this would introduce a four - cycle or a six - cycle in the graph . for any @xmath92 , @xmath95 only if @xmath96 . this implies that @xmath5 has two checks in @xmath97 . let @xmath81 be the set of such variables . we have the following lemma : there can be at most one variable in @xmath81 . suppose @xmath98 . specifically , assume @xmath99 . the proof is similar for @xmath100 . first note that for any @xmath101 , @xmath5 can not be connected to c14 as it would create a six - cycle . next , let @xmath102 and @xmath103 . then , @xmath5 can not have both checks in either c11 or c12 as this would cause a four - cycle . hence , @xmath5 has one check in c11 and one check in c12 . assume without loss of generality that v21 is connected to c11 and c16 . then , v22 can not be connected to c11 and c17 as this would form a six - cycle . v22 can not be connected to c12 and c17 as it would create a @xmath74 trapping set . hence , @xmath104 . let @xmath105 be connected to c11 , c16 and an additional check c21 . in the second iteration : @xmath106 we have the following lemma : [ lemma3:twoincorrectmessages ] there can not exist any variable @xmath107 such that it receives two or more incorrect messages at the end of the second iteration . suppose there existed a variable @xmath5 such that it received two incorrect messages in the second iteration . then , it would be connected to two checks in the set @xmath108 . this is not possible as it would introduce a four - cycle , six - cycle or a @xmath74 trapping set ( _ e.g. _ if @xmath5 is connected to c14 and c21 , it would form a @xmath74 trapping set ) . thus , in the third iteration : @xmath109 at the end of the third iteration , @xmath110 for all @xmath88 . also , we have the following lemma : there exists no @xmath92 such that @xmath111 . suppose there exists @xmath5 such that @xmath111 . then , @xmath5 is connected to three checks in the set @xmath112 . this implies that @xmath113 . however , from lemma [ lemma3:twoincorrectmessages ] it is evident that no such @xmath5 exists . hence , if a decision is made after the third iteration , a valid codeword is found and the decoder is successful . * subgraph 4 : * the variables in @xmath83 induce the subgraph shown in fig . [ config4 ] . at the end of the first iteration : @xmath114 for no @xmath115 , @xmath113 . for any @xmath115 , @xmath95 only if @xmath96 . let @xmath81 be the set of all such variables . we have the following lemma : \(i ) @xmath81 has at most four variables , and ( ii ) no two variables in @xmath81 can share a check in @xmath116 . _ sketch of the proof _ : there exists no variable which is connected to two checks from the set @xmath117 as it would introduce a four - cycle or a six - cycle . however , a variable node can be connected to one check from @xmath118 and to one check from @xmath119 . there can be at most four such variable nodes . when four such variable nodes exist , none are connected to c13 . also , these four variable nodes can not share checks outside the set @xmath120 . let these four variable nodes be labeled v21 , v22 , v23 and v24 and their third checks c21 , c22 , c23 and c24 , respectively . let @xmath121 . hence , in the second iteration : @xmath122 at the end of the second iteration @xmath123 for all @xmath88 . moreover , for no @xmath92 , @xmath113 . so , if a decision is made after the second iteration , a valid codeword is reached and the decoder is successful . * subgraph 5 : * the variables in @xmath83 induce the subgraph shown in fig . [ config5 ] . at the end of the first iteration : @xmath124 if there exists no variable @xmath115 such that @xmath93 , a valid codeword is reached after the first iteration . suppose this is not the case . let @xmath81 be the set of variables which receive two or more incorrect messages . then , we have the following lemma : \(i ) there exists one variable @xmath105 such that @xmath125 , and ( ii ) @xmath81 has at most three variables which receive two incorrect messages at the end of the first iteration . furthermore , they can not share a check in @xmath116 . we omit the proof of part ( ii ) as it is straightforward . part ( i ) is proved as follows : if there existed no variable , v21 , such that @xmath125 , then the decoder would converge in one iteration . next , suppose @xmath126 such that @xmath127 . without loss of generality , let v21 be connected to @xmath128 and @xmath129 . then , v21 would share two checks in the set @xmath79 . it is thennot possible to connect v22 without introducing a six - cycle or a @xmath74 trapping set ( _ e.g. _ , if v21 is connected to c12 , c15 and c18 , then it would introduce a @xmath74 trapping set ) . let the third checks connected to v22 , v23 and v24 be c21 , c22 and c23 , respectively and let @xmath130 in the second iteration : @xmath131 there can not exist a variable node which is connected to one check from @xmath132 and to one check from @xmath133 . also , there can not be a variable node which is connected to all three checks in the set @xmath132 as this would introduce a graph isomorphic to the @xmath73 trapping set . however , there can be at most two variable nodes which receive two incorrect messages from the checks in @xmath132 , say v31 and v32 . let the third checks connected to them be c31 and c32 , respectively . let @xmath134 and @xmath135 . at the end of the second iteration , variables v11 , v12 and v13 receive one incorrect message each . variables in the set @xmath136 receive two incorrect messages each . therefore , in the third iteration , we have : @xmath137 at the end of the third iteration , @xmath111 for all @xmath88 . furthermore , for no @xmath92 , @xmath111 . so , if a decision is made after the third iteration , a valid codeword is reached and the decoder is successful . in this section , we derive necessary and sufficient conditions for the correction of three errors in column - weight - four codes in four iterations of iterative decoding . this result is inspired by the analysis of error events in high - rate codes with column - weight four . in simulations , it was found that received vectors which did not converge to a valid codeword in the first 4 to 5 iterations did not converge thenceforth . hence , it is desirable to devise codes and decoding strategies in which vectors having a small number of errors converged rapidly to a codeword . to this end , it was found that a hybrid decoding strategy could correct three errors in four iterations if certain conditions are satisfied by the code . this result is summarized as follows : [ theorem:4:1 ] an ldpc code with column - weight four and girth six can correct three errors in four iterations of message - passing decoding if and only if the conditions , @xmath138 , @xmath139 , @xmath140 , @xmath141 and @xmath142 are satisfied . _ remark : _ it is worth noting that if a graph of girth six satisfies the @xmath138 condition , then it satisfies the @xmath139 condition as well . however , the addition of this extra constraint aids in the proof of the theorem . first , we prove the sufficiency of the conditions of theorem [ theorem:4:1 ] . let @xmath86 be the three erroneous variables . let @xmath79 be the set of checks that are connected to the variables in @xmath83 . the variables in @xmath83 can induce only one of the five subgraphs shown in fig . [ figure4 ] . we prove that in each case , the decoding algorithm converges to the correct codeword in four iterations . + + * subgraph 1 * : the variables in @xmath83 induce the subgraph shown in fig . [ figure4a ] . at the end of the first iteration , @xmath91 for all @xmath88 . moreover , no variable receives four incorrect messages after the first iteration as the existence of such a variable node would create a four - cycle . if a decision is made after the first iteration , the decoder is successful . * subgraph 2 * : the variables in @xmath83 induce the subgraph shown in fig . [ figure4b ] . at the end of the first iteration : @xmath143 for no @xmath115 , @xmath93 as it would introduce a four - cycle . for any @xmath92 , @xmath95 only if @xmath96 . this implies that @xmath5 is connected to three checks in @xmath144 . let @xmath81 denote the set of such variables . we have the following lemma : there can be at most three variables in @xmath81 . furthermore , no two variable nodes in @xmath81 share any check in the set @xmath116 . let @xmath145 . then the set of variable nodes @xmath146 has at most 15 neighboring checks . this violates the @xmath141 condition . hence , @xmath81 can have at most three variables . next , let @xmath126 . suppose they share a fourth check @xmath6 . since v13 can share at most two checks with v21 and v22 , assume that c110 and c111 are not neighbors of v21 , v22 . the neighbors of the variable nodes in the set @xmath147 all belong to the set @xmath148 which has cardinality 10 , thus violating the @xmath138 condition . let the fourth neighboring checks of v21 , v22 and v23 be c21 , c22 and c23 , respectively . let @xmath149 . in the second iteration : @xmath150 for all @xmath88 , @xmath123 . for no @xmath101 , @xmath113 . we now have the following lemma : there exists no variable @xmath107 such that @xmath113 . the proof is by contradiction . let @xmath107 such that @xmath113 . then , @xmath5 is connected to four checks in @xmath151 . note that only two neighbors of @xmath5 can belong to @xmath152 without introducing a four - cycle . this combined with the fact that there are at most three variable nodes in @xmath81 implies that there are only two cases : + ( a ) @xmath5 has two neighbors in @xmath152 and two neighbors in @xmath132 , say c21 and c22 . in this case , the set of variable nodes @xmath153 has @xmath154 check nodes , violating the @xmath140 condition . \(b ) @xmath5 has one neighbor in @xmath152 and three neighbors in @xmath132 . in this case , the set of variable nodes @xmath155 has @xmath156 check nodes , violating the @xmath141 condition . hence , if a decision is made after the second iteration , the decoder is successful . * subgraph 3 * : the variables in @xmath83 induce the subgraph shown in fig . [ figure4c ] . at the end of the first iteration , v11 , v12 and v13 receive correct messages from all their neighboring check nodes . moreover , there exists no variable which receives four incorrect messages from checks in the set @xmath79 . hence , if a decision is made after the first iteration , the decoder is successful . * subgraph 4 * : the variables in @xmath83 induce the subgraph shown in fig . [ figure4d ] . at the end of the first iteration : @xmath157 for no @xmath115 , @xmath93 as it this would introduce a four - cycle . for any @xmath115 , @xmath95 only if @xmath96 . this implies that @xmath5 has three checks in the set @xmath158 . let @xmath81 be the set of such variables . we now have the following lemma : there can be at most two variables in @xmath81 . moreover , there exists no check @xmath159 which is shared by two variables in the set @xmath81 . let @xmath160 . then , the set @xmath161 has at most @xmath154 checks which violates the @xmath162 condition . hence , @xmath81 has at most two variables . next , let two variables @xmath126 share a check @xmath159 . then , the set @xmath163 has at most 11 checks which violates the @xmath139 condition . let @xmath132 be the set of checks @xmath116 which are connected to variables in @xmath81 . in the second iteration we have : @xmath164 for no @xmath115 , @xmath113 , for such a structure can not exist without creating a four - cycle or violating one of @xmath139 and @xmath140 conditions . for any @xmath165 , @xmath166 only if @xmath167 . this implies that @xmath5 has three neighbors in the set @xmath168 . let @xmath136 be the set of such variables . we have the following lemma : for the sets @xmath81 and @xmath136 , the following are true : + ( i ) if @xmath169 , then @xmath136 is empty . + ( ii ) if @xmath170 , then @xmath171 . + ( iii ) @xmath172 . we prove the lemma part by part . + ( i ) suppose @xmath169 and that @xmath136 is not empty . let @xmath173 . then , the set @xmath174 is of size 6 and has at most @xmath154 checks which violates the @xmath140 condition . + ( ii ) suppose @xmath170 . let @xmath173 . if @xmath81 is empty , then v31 is connected to three checks in @xmath79 . this is not possible as @xmath175 . next , suppose that @xmath98 . let @xmath126 . then @xmath176 has at most @xmath154 checks which violates the @xmath140 condition . + ( iii ) suppose @xmath177 . by ( ii ) , there exists a variable @xmath105 . then , @xmath178 has at most @xmath154 checks which violates the @xmath140 condition . suppose @xmath179 . denote the fourth check of @xmath180 by @xmath181 . then , we have at the beginning of the third iteration : @xmath182 at the end of the fourth iteration , @xmath110 for @xmath88 . thus , if a decision is made at the end of this iteration , all @xmath88 are decoded correctly . now we prove the following lemma : there exists no @xmath183 such that @xmath111 . suppose that @xmath136 is empty and that there exists a variable @xmath5 such that @xmath111 . if @xmath81 is empty , then , @xmath5 is connected to four checks in @xmath184 . this is not possible as it would cause a four - cycle . if @xmath81 is not empty , then @xmath5 is connected to four checks in @xmath185 . then , we would have @xmath113 . however , from above , no such variable exists . next , suppose that @xmath186 and that @xmath187 . then , c31 is the only check such that @xmath188 and @xmath189 . it follows then that for any @xmath115 , if @xmath111 , then @xmath5 is connected to c31 . also , it is connected to three checks in the set @xmath190 . then the set of variables @xmath191 has at most @xmath192 checks . this violates the @xmath140 condition . hence , if a decision is made after the third iteration , the decoder is successful . * subgraph 5 * : the variables in @xmath83 induce the subgraph shown in fig . [ figure4e ] . for all @xmath88 , @xmath193 . there exist no @xmath115 that receive three incorrect messages , for the existence of such a variable would violate the @xmath138 condition . hence , for all @xmath115 , @xmath194 . let @xmath81 be the set of variables that have two checks in the set @xmath195 . let @xmath132 be the remaining two checks of these variables . at the beginning of the fourth iteration , the decoder switches to the gallager - b mode . then : @xmath196 at the end of the fourth iteration , @xmath197 for @xmath88 . moreover , for no @xmath115 , @xmath198 , as the existence of such a variable would either induce a four - cycle or violate one of the @xmath139 , @xmath140 , @xmath141 or @xmath142 conditions ( the arguments used are similar to the ones used for subgraph 2 and subgraph 4 ) . hence , if a decision is made at the end of the fourth iteration , the decoder is successful . now we prove the necessity of the conditions of the theorem . we prove this by giving subgraphs which violate _ one _ condition and are not successfully decoded in four iterations . since the validity of these claims can be checked easily , a detailed proof is omitted . * necessity of the @xmath199 condition * consider the subgraph shown in fig . [ figure4_10_subgraph ] . in this case , the @xmath138 condition is not satisfied and the errors are not corrected at the end of the fourth iteration . hence , in order to guarantee the correction of three errors in four iterations , the @xmath138 condition must be satisfied . subgraph[figure4_10_subgraph],scaledwidth=50.0% ] * necessity of the @xmath200 condition * there exists no graph of girth six which satisfies the @xmath138 condition but does not satisfy the @xmath139 condition . * necessity of the @xmath201 condition * consider the graph shown in fig . [ figure6_13_subgraph ] . the graph shown satisfies the @xmath138 and the @xmath139 conditions but not the @xmath140 condition . the errors are not corrected in four iterations . hence , in order to guarantee the correction of three errors in four iterations , the @xmath140 condition must be satisfied . subgraph[figure6_13_subgraph],scaledwidth=50.0% ] * necessity of the @xmath202 condition * consider the graph shown in fig . [ figure7_15_subgraph ] . the graph shown satisfies the probability @xmath138 , @xmath139 and the @xmath140 conditions but not the @xmath141 condition . the errors are not corrected at the end of the fourth iteration . hence , in order to guarantee the correction three errors in four iterations , the @xmath141 condition must be satisfied . subgraph[figure7_15_subgraph],scaledwidth=50.0% ] * necessity of the @xmath203 condition * consider the graph shown in fig . [ figure8_17_subgraph ] . the graph shown satisfies the @xmath138 , @xmath139 , @xmath140 and the @xmath141 condition but not the @xmath142 condition . the errors are not corrected at the end of the fourth iteration . hence , in order to guarantee the correction of three errors in four iterations , the @xmath142 condition must be satisfied . subgraph[figure8_17_subgraph],scaledwidth=50.0% ] in this section , we proved necessary and sufficient conditions to guarantee the correction of three errors in column - weight - four codes using an iterative decoding algorithm . by analyzing the messages being passed in subsequent iterations , it may be possible to get smaller bounds on the number of check nodes required in the `` small '' subgraphs however , we hypothesize that the size of subgraphs to be avoided would be larger . in this section , we describe a technique to construct codes with column - weight three and four which can correct three errors . codes capable of correcting a fixed number of errors show superior performance on the bsc at low values of transition probability @xmath204 . this is because the slope of the fer curve is related to the minimum critical number @xcite . a code which can correct @xmath45 errors has minimum critical number at least @xmath205 and the slope of the fer curve is @xmath205 . we restate the arguments from @xcite to make this connection clear . let @xmath204 be the transition probability of a bsc and @xmath206 be the number of configurations of received bits for which @xmath207 channel errors lead to codeword ( frame ) error . the frame error rate ( fer ) is given by : @xmath208 where @xmath45 is the minimal number of channel errors that can lead to a decoding error and @xmath21 is length of the code . on a semi - log scale the fer is given by @xmath209 for small @xmath204 , the expression above is dominated by the first two terms . that is , @xmath210 the @xmath211 vs. @xmath212 graph is close to a straight line with slope equal to @xmath213 the minimal critical number . if two codes @xmath214 and @xmath215 have minimum critical numbers @xmath216 and @xmath217 such that @xmath218 then the code @xmath214 will perform better than @xmath215 for small enough @xmath219 independent of the number of trapping sets . from the discussion in above and in section [ section3 ] , it is clear that for a code to have an fer curve with slope at least @xmath220 , the corresponding tanner graph should not contain the trapping sets shown in fig . [ trappingsets ] as subgraphs . we now describe a method to construct such codes . the method can be seen as a modification of the peg construction technique used by hu _ the algorithm is detailed below as algorithm [ algorithm:1 ] . [ algorithm:1 ] note that checking for a graph isomorphic to @xmath73 trapping set is computationally complex . since , the peg construction empirically gives good codes , it is unlikely that it introduces a weight - eight codeword . however , once the graph is grown fully , it can be checked for the presence of weight - eight codewords and these can be removed by swapping few edges . using the above algorithm , a column - weight - three code with @xmath221 variable nodes and @xmath222 check nodes was constructed . the code has slight irregularity in check degree . there is one check node degree five and one check node with degree seven , but the remaining have degree six . the code has rate 0.5 . in the algorithm , we restrict maximum check degree to seven . the performance of the code on bsc is compared with the peg code of same length . the peg code is empirically the best known code at that length on awgn channel @xcite . however , it has fourteen @xmath72 trapping sets . [ pegnewvsold ] shows the performance comparison of the two codes . as can be seen , the new code performs better than the original peg code at small values of @xmath204 . , scaledwidth=60.0% ] unlike column - weight - three codes , the construction of column - weight - four codes involves ensuring certain expansion on subsets of variable nodes . this can be done only in time which grows exponentially with the length of the code . hence , we consider the @xmath223 condition rather than the necessary and sufficient conditions discussed in section [ section4 ] . it can be shown that the @xmath223 condition is sufficient for the @xmath138 , @xmath139 , @xmath140 , @xmath141 and the @xmath142 conditions . there are only two graphs of girth @xmath224 with @xmath220 variable nodes and @xmath225 check nodes [ figure:4_11 ] shows these two graphs . avoiding these two subgraphs will ensure a code which can correct three errors . an algorithm for the construction of such codes is similar to the modified peg algorithm given in algorithm [ algorithm:1 ] . this algorithm was used to generate a code of length @xmath226 , girth @xmath224 and rate @xmath227 . the code constructed has a slight irregularity in that three check nodes have degree nine and three have degree seven . _ remark : _ for the code parameters given above , it was possible to generate a code which satisfied the @xmath223 condition . however , it might not be possible to satisfy this condition for codes with higher rate and/or shorter lengths . should such a scenario arise , the set of subgraphs to be avoided should be changed ( _ e.g. _ , to those specified in the necessary and sufficient conditions ) . however , the code construction time will be larger . hence , at the cost of code - construction time and complexity , it is possible to achieve shorter lengths and/or higher rates . [ figure:4_11_subgraph_1 ] variable nodes and @xmath225 check nodes . subgraphs with @xmath220 variable nodes and fewer than @xmath225 check nodes do not exist.[figure:4_11 ] , title="fig:",scaledwidth=40.0% ] [ figure:4_11_subgraph_2 ] variable nodes and @xmath225 check nodes . subgraphs with @xmath220 variable nodes and fewer than @xmath225 check nodes do not exist.[figure:4_11 ] , title="fig:",scaledwidth=40.0% ] fig . [ figure : colwtfourplot ] shows the performance of the code under message - passing decoding . the curve on the left corresponds to four iterations of message - passing . the curve in the right corresponds to @xmath228 iterations of message - passing . after only four iterations , errors of weight four and above were encountered which were not corrected by the message - passing decoder . however , after @xmath228 iterations , the smallest weight error pattern still remaining had a weight of @xmath229 . we note that the average slope of the fer curve is @xmath230 which is the weight of the dominant error event at these probabilities of error . this suggests that analysis over a higher number of iterations and on `` larger '' subgraph search will yield a stronger result . however , this is beyond the scope of this paper . also , it is worth noting that the conditions of theorem [ theorem:4:1 ] avoid codewords of length @xmath220 through @xmath230 which improves the minimum distance of the code . , in this paper , we provided a method to derive conditions that guarantee the correction of a finite number of errors by hard - decision decoding . although more involved than the expander arguments used in previous works , it results in better bounds . moreover , in contrast to previous expansion arguments , our results give rise to code - construction techniques that yield codes with guaranteed error - correction ability under message - massing decoding at _ practically feasible lengths_. this method can be applied to ( a ) provide conditions for guaranteed correction of a larger number of errors , ( b ) yield similar results for higher column - weights and/or higher girths . however , such applications would be more involved than the analysis done in this work . d. j. c. mackay and m. j. postol , `` weaknesses of margulis and ramanujan margulis low - density parity - check codes , '' in _ proc . of mfcsit2002 , galway _ , ser . electronic notes in theoretical computer science , vol . 74.1em plus 0.5em minus 0.4emelsevier , 2003 . [ online ] . available : http://www.inference.phy.cam.ac.uk/mackay/abstracts/margulis.html s. k. chilappagari , s. sankaranarayanan , and b. vasic , `` error floors of ldpc codes on the binary symmetric channel , '' in _ proc . of international conference on communications _ , vol . 3 , june 11 - 15 2006 , pp . 10891094 . y. kuo , s. lin , and m. p. c. fossorier , `` low - density parity check codes based on finite geometries : a rediscovery and new results , '' _ ieee trans . inform . theory _ 47 , no . 7 , pp . 27112736 , nov . 2001 . m. ivkovic , s. k. chilappagari , and b. vasic , `` eliminating trapping sets in low - density parity - check codes by using tanner graph covers , '' _ ieee trans . inf . theory _ , 54 , no . 8 , pp . 37633768 , 2008 . [ online ] . available : http://dx.doi.org/10.1109/tit.2008.926319	in this paper , we give necessary and sufficient conditions for low - density parity - check ( ldpc ) codes with column - weight three to correct three errors when decoded using hard - decision message - passing decoding . additionally , we give necessary and sufficient conditions for column - weight - four codes to correct three errors in four iterations of hard - decision message - passing decoding . we then give a construction technique which results in codes satisfying these conditions . we also provide numerical assessment of code performance via simulation results . submitted to ieee transactions on information theory , october 2008
starburst galaxies can be grouped in two main families : the hii galaxies and the starburst nucleus galaxies ( sbngs ) . the former are small - mass and metal - poor galaxies , while the latter are more massive and metal - rich ( for a complete definition of starburst galaxies , see coziol et al . the astronomical community recognized quite early the peculiar nature of h ii galaxies and many studies have been undertaken to determine their characteristics . the origin of sbngs , on the other hand , was thought to be more straightforward to understand . as a consequence , the basic characteristics of these galaxies are not as well determined and their nature is still in debate . for example , it was believed some twenty years ago that sbngs were `` old '' galaxies which were rejuvenated by inflow of matter resulting from interactions with other galaxies ( huchra 1977 ; tinsley & larson 1978 ) . but subsequent studies have shown that starbursts generally do not reside in high galaxy density regions ( salzer et al . 1989 ; hashimoto et al . 1998 ; coziol et al . 1998b , 1999 ) , where conditions for interaction should be more favorable , and that only a fraction ( between 25% and 30% ) of the massive sbngs have obvious luminous companions ( keel & van soest 1992 ; gallimore & keel 1993 ; coziol et al . 1995 ; coziol et al . 1997a , 1999 ; contini et al . 1998 ) . the frequency of interacting galaxies among sbngs may be higher however , if we assume that the morphological peculiarities often observed in these galaxies , are weak traces of past interactions ( keel & van soest 1992 ; coziol et al . 1995 , barth et al . 1995 ) and that , consequently , many sbngs are the remnants of merging galaxies ( coziol et al . the assumption that sbngs result from accretions of smaller galaxies is supported by several facts : they are less chemically evolved than normal galaxies with similar morphologies and comparable luminosities ( coziol et al . 1997b ) , they are predominantly early type spiral galaxies ( coziol et al . 1997a ; 1998a ) and follow a luminosity metallicity relation similar to that of elliptical galaxies ( coziol et al . 1998c ) . the merger hypothesis , on the other hand , does not completely explain the starburst phenomenon and it is probably also necessary to assume internal regulating star formation mechanisms , in order to extend the total lifetime of the starburst over a longer period than is usually assumed based on the dynamical time scale of a gravitational interaction between galaxies ( searle & sargent 1972 ; gerola et al . 1980 ; krgel & tutukov 1993 ) . such mechanisms would explain why sbngs experienced more than one burst over the last few gyrs ( gonzlez delgado et al . 1995 ; coziol 1996 ) . they would also explain why these galaxies do not produce supernovae at higher rates than normal ones ( turatto et al . 1989 ; richmond et al . 1998 ; gonzlez delgado et al . 1999 ) in principle , one could gain a better understanding of the nature and origin of sbngs by drawing a more complete picture of their chemical evolution . but this is a difficult task , which is further complicated by our relative ignorance of how the chemical evolution of normal spiral galaxies proceeds . the main difficulty is observational : building a reasonable scenario of the chemical evolution of galaxies generally requires observations of hii regions over a large range of wavelengths . for local galaxies , for example , uv observations are required for determining the abundance of carbon ( kobulnicky & skillman 1998 ) , while observations in the near infrared are needed for estimating the abundance of sulphur ( garnett 1989 ) . in metal - rich spiral galaxies , an additional difficulty resides in the non observability of important nebular lines , like [ oiii ] at 4363 , without which the electron temperature of the gas can not be determined directly . many efforts have been devoted over the years to develop functional empirical techniques to circumvent the problems related to the determination of chemical abundances in galaxies ( pagel et al . 1979 ; edmunds & pagel 1984 ; edmunds 1989 ; skillman 1989 ) . recently , thurston et al . ( 1996 ; hereafter teh ) have extended these techniques and devised a new method for estimating nitrogen abundances in metal - rich galaxies . in this paper , we take advantage of this recent advance to determine the abundance of nitrogen in sbngs . the organization of our paper is as follows . in section 2 , we identify the main source of excitation in our sample of sbngs , based on their location in three standard diagnostic diagrams . we then discuss , in section 3 , the legitimacy of applying to sbngs the empirical method designed for `` normal '' hii regions to determine their chemical abundances . in the sbngs , we took great care in verifying that no other sources of excitation like a hidden agn is present , which could complicate or even impede the abundance determination in these galaxies . having verified these points , we proceed in section 4 to estimate the oxygen and nitrogen abundances in a large sample of sbngs , comparing these abundances with those of normal hii regions in spiral galaxies . in section 5 , we discuss two different scenarios for producing nitrogen in sbngs and examine their consequences for the origin of the bursts in these galaxies . a summary of our most salient results is presented in section 6 . the optical spectrum of a starburst galaxy is usually quite distinct from that of an agn . it resembles that of an hii region in our galaxy , where , as we know , the gas is photoionized by young o and b stars ( searle & sargent 1972 ; sargent 1972 ; huchra 1977 ; french 1980 ) . this conclusion applies to all the types of starburst galaxies known today ( salzer et al . 1989 ; kim et al . 1998 ; coziol 1996 ; coziol et al . 1998a ) , including those where the burst is located mostly in the central regions ( weedman et al . 1981 ; balzano 1983 ) . but , because the physical conditions in the nuclei of galaxies are still poorly understood , one may wonder what kind of pathological phenomena the evolution of a powerful starburst in the center of a galaxy could develop ( weedman 1983 ; terlevich et al . 1991 ) . this concern may partly be justified by observations . for example , the presence of low - intensity non - stellar activity in the center of normal galaxies is frequently suggested ( heckman 1980 ; peimbert & torres - peimbert 1981 ; vron et al . 1981 ; rose & cecil 1983 ; filippenko & sargent 1985 ) and , in agns , traces of star formation occurring simultaneously with an active nucleus are sometimes noted ( shields & filippenko 1990 ; lonsdale et al . 1993 ; vron et al . 1997 ; gonzlez delgado et al . however , none of these observations constitutes definite evidence that an evolutionary connection between starburst and agn exists , and , even less so , explains how this evolution may proceed . -2 cm -2.cm obviously , if a source of excitation other than star formation is active in sbngs , this will greatly complicate abundance determinations . it is thus essential to first establish the dominant excitation source ( o and b stars or an agn ) present in these galaxies . this is usually done empirically : one constructs diagnostic diagrams using several ratios of emission lines present in the optical spectra ( baldwin et al . 1981 ; veilleux & osterbrock 1987 , hereafter bpt and vo ) . over the years , the use of diagnostic diagrams has been recognized as the most efficient method for classifying the different types of activity encountered in emission - line galaxies ( veilleux et al . 1995 ; vron et al . 1997 ; gonalves et al . 1999 ) . this is the method that we applied to identify the main source of excitation in our different samples of galaxies . our sample of uv - bright sbngs was composed originally of 208 hii regions observed along the bars of 75 markarian barred galaxies ( see contini 1996 , contini et al . 1998 and considre et al . 1999 for details ) . from this sample , we rejected 48 hii regions because of their ambiguous classification in the three diagnostic diagrams . this same criterion was used by veilleux et al . ( 1995 ) to build their sample of fir - bright sbngs . this criterion allows one to eliminate the so called `` transition '' galaxies , whose nature is not well determined ( vron et al . 1997 ; gonalves et al . 1999 ; hill et al . 1999 ; doyon et al . 1999 ) . for comparison , we used the sample of 83 fir - bright sbngs of veilleux et al . we also used the 70 hii regions observed by mccall et al . ( 1985 ; hereafter mrs ) to represent normal hii regions . in all the samples , the line ratios were corrected for internal reddening using balmer decrements . the three classical diagrams of [ oiii]@xmath2/h@xmath3 ( the excitation level ) as a function of [ nii]@xmath4/h@xmath5 , [ sii]@xmath6/h@xmath5 and [ o i]@xmath7/h@xmath5 for the various samples are shown in fig . [ fig.1 ] . note that , for the sake of clarity , these line ratios will hereafter be written without the wavelengths . agns ( seyfert 2s and liners ) are clearly separated from hii regions and sbngs ( fig . [ fig.1]a , d and g ) . in these diagrams , the hii regions trace a sequence where the decreasing excitation is due to the increasing oxygen abundance ( mrs ; dopita & evans 1986 ) . in general , the sbngs trace a similar sequence , but slightly displaced towards higher emission - line ratios ( coziol et al . 1997a ) . despite being slightly more luminous in the infrared , the fir - bright sbngs show optical spectral characteristics similar to those of the uv - bright sbngs . this is consistent with the similarities observed between these two types of sbngs in the fir and suggests that the two selection criteria ( uv- or fir - bright ) define mostly the same kind of galaxies ( coziol et al . 1998a ) . a kolmogorov smirnoff test performed on [ nii ] suggests , with a confidence limit greater than 99% , that the fir- and uv bright sbngs come from the same population . the same test yields similar conclusions for [ oi ] and [ sii ] , but only with confidence limits greater than 90% and 80% respectively . the [ oi ] line is rarely observed in normal hii regions ( bpt ) , and it was not measured in the mrs sample . when it is measurable in starburst , its intensity is always weaker than in agns . this is because the [ oi ] line is produced only in regions of partial ionization which are much more extended in agns ( vo ) . in our sample of uv bright sbngs , only a small fraction ( @xmath8 30% ) of the galaxies presents this line . we compare the few [ oi]/h@xmath5 ratios observed in the sbngs of our sample in fig . [ fig.1]h and i. log([oi]/h@xmath5 ) is on average @xmath9 , comparable to what is observed in the sample of fir - bright sbngs of veilleux et al . ( 1995 ) and consistent with values found in normal hii regions ( bpt , vo ) . we conclude , on the basis of the three standard diagnostic diagrams , that the main source of excitation of the gas in sbngs is o and b stars , like in normal hii regions ( bpt ; vo ; gonalvez et al . 1998 ; kim et al . 1998 ) . as can be seen in fig . [ fig.1 ] , even if the gas in sbngs is ionized by o and b stars , the spectra of these galaxies do show , however , an excess of nitrogen and sulphur emissions with respect to hii regions in the disk of normal spirals ( the mrs sample ) . before proceeding with the determination of abundances in sbngs , we will now discuss the possible causes of this phenomenon . note that because we do not have enough information to estimate the abundance of sulphur in our samples , our discussion will focus on nitrogen . an excess of nitrogen emission in the spectra of galaxies ( usually normal ones ) has been observed before ( see stauffer 1982 , and references therein ) and various hypotheses have been proposed to explain it ( e.g. kennicutt et al . 1989 ) . the simplest hypothesis , which can apply to sbngs , is that these galaxies possess an overabundance of nitrogen with respect to normal disk hii regions ( stauffer 1982 ) . such an overabundance has already been observed in the prototypical sbng ngc 7714 ( gonzlez delgado et al . 1995 ) and some have suggested that it also happens in agns ( storchi bergmann 1991 ; storchi - bergmann & wilson 1996 ; ohyama et al . 1997 ) . another possibility is that sbngs have an extra source of excitation . this , in fact , was the solution preferred by kennicutt et al . ( 1989 ) , who suggested a hidden agn as the extra source . as proposed by lehnert & heckman ( 1994 ) , shocks produced by supernovae or starburst winds could probably also do the trick . however , the extra excitation hypothesis is in contradiction with the use of emission line diagnostic diagrams to classify activity types in galaxies . indeed , we remind the reader that we already eliminated from our sample all the cases which appeared ambiguous in these diagnostic diagrams . therefore , if the excess emission observed in these galaxies is still due to a residual excitation effect , it would seriously question the use of these diagnostic diagrams for distinguishing between the different activity types in emission line galaxies . at the present time however , there is no evidence that diagnostic diagrams can give a wrong classification ( bpt ; vo ; gonalvez et al . 1998 ; kim et al . 1998 ) . as an alternative , shields & kennicutt ( 1995 ) have suggested that metal depletion occurring during dust formation could produce a slight extra excitation . at first , this seems perfectly applicable to the fir - bright sbngs . however , the fact that we do not observe any difference in the diagnostic diagrams between these dusty galaxies and their uv - bright counterparts suggests that dust does not play a significant role in producing the excess of emission in these galaxies . this interpretation is consistent with the original analysis made by vo on the null effect of dust on the different line ratios used in the diagnostic diagrams . it is also consistent with the fact that the two types of sbngs have similar properties in the far - infrared ( coziol et al . 1998a ) . if the excess emission in the sbngs with respect to normal disk hii regions is due to a residual excitation effect , we can not use the empirical method devised by teh to determine the abundance of nitrogen in sbngs . in fact , we would even caution against using this method to determine the nitrogen abundance , as teh have done , in early - type spirals and the nuclei of normal galaxies . it is therefore important to test thoroughly the different hypotheses for the origin of the excess emission . @xmath10/h}\alpha ) & \log({\rm [ oiii]/h}\beta ) & \delta({\rm [ nii]/h}\alpha)\\ \noalign{\smallskip } \hline -1.0 & 22 & -0.37\pm0.12 & -0.95\pm0.09 & + 0.19 \\ -0.6 & 87 & -0.39\pm0.11 & -0.58\pm0.10 & + 0.09 \\ -0.2 & 72 & -0.42\pm0.11 & -0.21\pm0.12 & + 0.16 \\ + 0.2 & 52 & -0.68\pm0.20 & + 0.19\pm0.11 & + 0.20 \\ + 0.6 & 10 & -1.05\pm0.16 & + 0.53\pm0.08 & + 0.21 \\ \noalign{\smallskip } \hline \noalign{\smallskip } { \rm center\ bin } & { \rm n } & \log({\rm [ sii]/h}\alpha ) & \log({\rm [ oiii]/h}\beta ) & \delta({\rm [ sii]/h}\alpha)\\ \noalign{\smallskip } \hline -1.0 & 22 & -0.63\pm0.11 & -0.94\pm0.08 & + 0.15 \\ -0.6 & 87 & -0.60\pm0.10 & -0.58\pm0.11 & + 0.07 \\ -0.2 & 72 & -0.57\pm0.09 & -0.21\pm0.12 & + 0.13 \\ + 0.2 & 52 & -0.58\pm0.14 & + 0.19\pm0.11 & + 0.21 \\ + 0.6 & 10 & -0.72\pm0.14 & + 0.53\pm0.12 & + 0.25 \\ \noalign{\smallskip } \hline \noalign{\smallskip } { \rm center\ bin } & { \rm n } & \log({\rm [ oi]/h}\alpha ) & \log({\rm [ oiii]/h}\beta ) & ---\\ \noalign{\smallskip } \hline -1.0 & 9 & -1.60\pm0.18 & -0.91\pm0.06 & ... \\ -0.6 & 49 & -1.60\pm0.27 & -0.57\pm0.10 & ... \\ -0.2 & 49 & -1.45\pm0.18 & -0.22\pm0.12 & ... \\ + 0.2 & 21 & -1.56\pm0.16 & + 0.18\pm0.11 & ... \\ + 0.6 & 5 & -1.61\pm0.14 & + 0.52\pm0.08 & ... \\ \noalign{\smallskip } \hline \end{array}\ ] ] if the excess emission observed in sbngs is caused by a surplus of excitation , the intensities of the emission lines should be correlated ; they should all increase simultaneously . to look for such a correlation , we have computed the mean values of [ nii ] , [ sii ] and [ oi ] ( relative to h@xmath5 ) in excitation bins of 0.4 dex . these mean values are given in table 1 and plotted in fig [ fig.1]c , f and i. because we did not find any difference between the spectral characteristics of the fir- and uv - bright sbngs , we merged the two samples together for this analysis . the mean excess of emission with respect to normal disk hii regions is about the same for nitrogen and sulphur , @xmath11 dex and @xmath12 dex respectively however , the difference in [ nii]/h@xmath5 stays almost constant while the difference in [ sii]/h@xmath5 increases with the level of excitation ( last column of table 1 ) . it also appears that [ nii]/h@xmath5 increases as the level of excitation decreases ( or as the oxygen abundance increases ) , while [ s ii]/h@xmath5 and [ oi]/h@xmath5 stay constant . in fig [ fig.1]c , f and i , it can be seen that [ nii]/h@xmath5 is correlated with [ o iii]/h@xmath3 , while [ sii]/h@xmath5 and [ oi]/h@xmath5 are not . this is confirmed with a confidence limit higher than 99% using a generalized kendall s tau correlation coefficient test . these different behaviors of the emission - line ratios with the level of excitation are inconsistent with the hypothesis that the excess of emission in sbngs is produced by an additional excitation effect . on the other hand , they are partly consistent with what is expected if these line ratios are sensitive to a variation of abundance . if nitrogen is produced by intermediate - mass stars while sulphur is produced by massive stars , the intensities of these two lines should not be correlated . the different behaviors of these two lines also suggest that , in sbngs , the abundance of nitrogen will increase with that of oxygen while the abundance of sulphur will stay constant , as is observed in normal disk h ii regions ( torres peimbert et al . we will show later ( sect.5.1 ) that these behaviors are consistent with a scenario of successive bursts for the chemical evolution of sbngs . to better understand the difference between sbngs and agns , we compare [ nii]/h@xmath5 with [ sii]/h@xmath5 in our different samples . the two line ratios are significantly higher in the liners and seyfert 2s than in the sbngs ( see fig . [ fig.2]a and b ) . note that we can not distinguish between the two agn types in this diagram . the sbngs , on the other hand , are clearly separated from liners . some seyfert 2s have [ sii]/h@xmath5 ratios comparable to those of sbngs , but this is not a problem because seyfert 2s can be clearly distinguished from sbngs based on their higher excitation level in fig . [ fig.1 ] . -2 cm -4 cm using the positions of the sbngs and seyfert 2s / liners in fig . [ fig.2 ] , we establish two regions in this diagram where the gas is excited by different mechanisms : agn vs photoionization by o and b stars . this separation is also consistent with the lower limits proposed by lehnert & heckman ( 1996 ) for the presence of diffuse ionized gas in the halos of edge - on starbursts . we test our separation criterion on the uv - bright sbngs and on the mrs sample of disk hii regions in fig . [ fig.2]c and d. as expected , all the hii regions in the mrs sample and the uv - bright sbngs are located in the region where excitation by o and b stars is predominant . we also show in fig . [ fig.2]d the position of the hii regions in our sample of sbngs which were rejected on the basis of an ambiguous classification in the three diagnostic diagrams . most of these ambiguous cases are either in the agn regions or at the border of these regions . this supports our conclusion that the main source of excitation of the gas in sbngs is o and b stars . for our analysis to be complete , we now verify that no other factors influence the line ratios under investigation . the sulphur lines , for example , are sensitive indicators of electronic density , and one could imagine that a variation of this density over the line of sight introduces spurious effects . however , as shown in fig . [ fig.3]a , the three line ratios are independent of electronic density . -2 cm -6.1 cm another spurious effect will arise if the slit of the spectrograph covers regions with different physical properties , e.g. a central agn and a circumnuclear starburst region , producing intermediate values for the line ratios . the presence of such an effect can be tested by its dependence on redshift . if the physical conditions ( different sources of excitation or density layers ) change with radial distance , while the slit aperture remains constant , the line ratios will change with redshift . for example , an agn in the nucleus of each sbng would produce higher line ratios at low redshifts , because of the increasing contribution of the agn over that of circumnuclear hii regions ( storchi - bergmann 1991 ) . the contrary is expected for diffuse gas ( lehnert & heckman 1994 , 1996 ) ; the ratios would then increase with redshift due to the enhanced contribution of diffuse gas in the aperture . none of these effects is observed in fig . [ fig.3]b . we also checked if the line ratios vary with the inclination of the galaxies ( as estimated by their axis ratio _ b / a _ ) ; this is shown in fig . [ fig.3]c . lehnert & heckman ( 1996 ) found different variations of the line ratios along the two axes , which they attributed to the presence of diffuse ionized gas in the halos of some fir - bright starbursts . again , no significant effect is present in the two samples of sbngs . @xmath13 } & 3727 & 1.235 & 1.236 & 1.752 \\ { \rm [ neiii ] } & 3869 & 0.264 & 0.259 & 0.350 \\ { \rm h}\gamma & 4340 & 0.412 & 0.404 & 0.476 \\ { \rm [ oiii ] } & 4363 & 0.028 & 0.031 & 0.036 \\ { \rm h}\beta & 4861 & 1.000 & 1.000 & 1.000 \\ { \rm [ oiii ] } & 4959 & 1.427 & 1.483 & 1.456 \\ { \rm hei } & 5876 & 0.142 & 0.142 & 0.114 \\ { \rm h}\alpha & 6563 & 3.997 & 3.980 & 2.833 \\ { \rm [ nii ] } & 6584 & 0.206 & 0.206 & 0.146 \\ { \rm [ sii ] } & 6716 & 0.175 & 0.173 & 0.120 \\ { \rm [ sii ] } & 6731 & 0.132 & 0.131 & 0.091 \\ { \rm [ ariii ] } & 7135 & 0.177 & 0.176 & 0.116 \\ { \rm [ oii ] } & 7325 & 0.101 & 0.061 & 0.039 \\ \noalign{\smallskip}\hline \end{array}\ ] ] unreddened line intensities using the curve for lmc of pei ( 1992 ) finally , we looked for a relation between the emission - line ratios and the intensity of the burst , as measured by the ratio l@xmath14/l@xmath15 . following lehnert & heckman ( 1996 ) , the most straightforward prediction of the superwind theory is a relation between the emission - line ratios and the star formation rates . no relation exists between the three line ratios and the intensities of the bursts in the fir - bright sbngs ; this is shown in fig . [ fig.3]d . the [ nii]/h@xmath5 ratio increases slightly with l@xmath14/l@xmath15 in the uv - bright sbngs , but this is because [ n ii]/h@xmath5 is lower in the sbngs with high excitation , which also happen to have the lowest star formation rates . all these negative results indicate that no spurious effects invalidate the results of our analysis of the emission - line ratios in sbngs . we thus conclude that there is no extra excitation effect in these galaxies . we are therefore left with only one possible explanation for the origin of the excess of nitrogen emission in sbngs , namely that they have different abundances than hii regions in normal spiral galaxies ( stauffer 1982 ) . this is what we will now verify . in their analysis of photoionized models of hii regions , evans & dopita ( 1985 ) pointed out that the position occupied by a galaxy in the diagnostic diagram [ oiii]/h@xmath3 vs [ nii]/h@xmath5 should also depend on the ratio n / o . it seems therefore reasonable to think that the higher [ nii]/h@xmath5 ratios in the sbngs are due to an abundance effect . indeed , a higher dispersion in n / o for a given metallicity will automatically translate into a higher dispersion of [ nii]/h@xmath5 for a given level of excitation , that is for a fixed [ oiii]/h@xmath3 . to test this hypothesis we constructed two different hii region sequences using the photoionization code cloudy ( ferland 1997 ) . for temperature @xmath16 k , we used kurucz atmospheres , while for higher temperature , the non lte model of rauch ( 1997 ) was adopted . the resulting sequences are shown in fig.[fig.4 ] . our first sequence ( seq . 1 ) was built to reproduce the spectral sequence for normal hii regions represented by the mrs sample . to constrain the sequence at the low - metallicity end , we used the parameters ( table 2 ) and abundances ( table 3 ) of the lmc hii region n159 ( observed by dufour et al . 1982 ) , deduced from a previous model also based on cloudy ( carlos reyes et al . 1998 ) . this reference point is crucial because the models of mrs and dopita & evans ( 1986 ) do not reproduce the observations in this part of the diagnostic diagram . at a low metallicity level , we assumed that hii regions are photoionized by very young and hot stars . @xmath17 @xmath18 carlos reyes et al . ( 1998 ) grevesse & noels ( 1993 ) starting from this point , we reproduced the spectral sequence for normal hii regions by varying the physical parameters and chemical abundances . for each abundance value , we searched for the combination of luminosity and ionization temperature that best reproduces the spectral sequence of hii regions . except for he , c and n , the abundances changed linearly . the abundance of he was kept constant , while the ratios c / o and n / o varied following eqs . ( 1 ) and ( 2 ) . @xmath19 \nonumber \\ & - 0.49[\log({\rm o / h } ) + 12]^2 \end{aligned}\ ] ] @xmath20 \nonumber \\ & + 0.26[\log({\rm o / h } ) + 12]^2 \end{aligned}\ ] ] these two parametric relations were obtained from a regression on data from the lmc , smc and solar values ( table 4 ) . ( 2 ) yields a n / o ratio for the lmc which is consistent with the value given by the @xmath21 relation proposed by vila costas & edmunds ( 1993 ; hereafter vc&e ) . in the diagnostic diagram , the position of the galaxies with solar abundance was determined using the metallicity calibration curve of coziol et al . ( 1994 ) . -3.8 cm -9.5 cm sequence 2 was derived from sequence 1 by increasing by a factor 2 the abundance of nitrogen predicted by eq . ( 2 ) . as can be seen in fig.[fig.4 ] , sequence 2 reproduces the distribution in sbngs relatively well . note that the metallicities are comparable to those found in sequence 1 . therefore , for a given metallicity , this sequence predicts an overabundance of nitrogen in sbngs with respect to normal disk hii regions . for comparison , we have also drawn in fig.[fig.4 ] the predictions made by the model of shields & kennicutt ( 1995 ) which takes into account the depletion of metals in dust . this model generally overpredicts the excess of nitrogen emission in sbngs . considering the relatively good fit of sequence 2 , we conclude that the excess of nitrogen emission in sbngs can easily be reproduced by a slight overabundance of this element . in the previous section , we performed several tests to verify the legitimacy of applying to the sbngs the method for determining the nebular abundance in normal hii regions . first , we identified the main source of excitation of the gas in sbngs with o and b stars . then by constructing a simple sequence of hii regions , we showed that the simplest hypothesis for explaining the excess of nitrogen emission observed in these galaxies is a slight overabundance of this element with respect to normal disk hii regions . we now feel justified to determine the oxygen and nitrogen abundances in our sample of sbngs with the empirical techniques developed for normal hii regions . in doing so , we should also verify the predictions of our photoionized model , by comparing the abundances in the sbngs with those of normal disk hii regions . the oxygen abundances are determined following the relation ( vacca & conti 1992 ) : @xmath22 where r@xmath23([oiii]@xmath24 . to determine the n / o ratios , we use the teh method , adopting their empirical equation to estimate the electron temperature @xmath25 in the [ nii ] emission region . using simple algorithms based on the three - level - atom solutions of mccall ( 1984 ) , the abundance ratio n / o , assumed to be equal to n@xmath26/o@xmath26 , is determined following the relation : @xmath27 where @xmath25 is in units of @xmath28 k. for some of the galaxies in our sample , [ oii]@xmath29 was estimated using the empirical relation between [ o ii]@xmath30h@xmath3 and [ oiii]@xmath31h@xmath3 proposed by mrs . using their sample of galaxies , we verified that the n / o ratios obtained using the observed and estimated values of [ oii ] are the same . we also tested this relation for the sbngs using the galaxies in our sample where [ oii]@xmath29 was directly measured . -4.5 cm -9.5 cm we compare in fig . [ fig.5 ] the predictions of sequences 1 and 2 with the n / o estimates for the mrs sample of normal hii regions as well as for a subset of hii regions from the compilation of vc&e . the n / o estimates for the mrs sample of galaxies are @xmath32 dex lower , on average , than the values predicted by sequence 1 . in fig . [ fig.5 ] , we also plotted the fits of the different models ( @xmath33 and @xmath21 production of nitrogen ) proposed by vc&e to explain the values observed in different normal hii regions . in general , the model for @xmath33 production seems to better describe the behavior of normal hii regions , but it underestimates the n / o ratio at solar oxygen abundance . -4.5 cm -9.5 cm we verified that the offset of 0.15 dex between the data and sequence 1 does not depend on the relation adopted for estimating the oxygen abundance . if we use the empirical equation proposed by zaritsky et al . ( 1994 ) for example ( or a similar relation deduced from our model ) the oxygen abundances of the most metal - rich hii regions are slightly reduced , changing the slope of the mrs distribution and rendering it more similar to the @xmath33 relation . we could also add a term in the equation giving the n / o ratio . indeed , the complete equation ( see pagel et al . 1992 , their eq . 9 ) contains another term which depends on density and temperature . but this additional term is two orders too low to explain the offset , even if we assume extreme physical conditions ( temperatures lower than @xmath34 k and densities well over 1000 @xmath35 ) . -0.5 cm -8 cm although 0.15 dex is probably smaller than the uncertainties on our estimates , we believe that the offset of the mrs sample with respect to sequence 1 is real . in fact , mrs already observed such a difference with their own models . in support of this hypothesis , we note that the 158 disk hii regions from the compilation of vc&e in general follow the @xmath33 relation in fig . [ fig.5 ] . the same result was found recently by van zee et al . it seems , therefore , that normal hii regions usually have lower than solar n / o values at solar metallicity . consequently , the behavior of the hii regions in the mrs sample is normal . our estimates of the n / o ratios in sbngs are shown in fig . [ fig.6 ] . in these galaxies , a clear excess of nitrogen with respect to normal hii regions is observed . for the sake of comparison , the predictions of our sequences 1 and 2 were plotted in fig . [ fig.6 ] with an offset of @xmath36 dex in n / o . from this , we deduce that the abundance of nitrogen in the sbngs is at most a factor two higher than in normal disk hii regions . examining how the abundance of nitrogen varies with that of oxygen , we find that the sbngs do not follow the @xmath33 relation . this is shown in fig . [ fig.6 ] . a linear fit yields ( with a correlation coefficient of 76% ) @xmath37n / o@xmath38o / h@xmath39 , which is consistent with a mixture of @xmath0 mode of production ( mcgaugh 1991 ) . but the increase of the n / o ratio with metallicity does not look like a continuous process . the n / o ratio rises sharply by about 0.3 dex at an oxygen abundance of @xmath40 and stays almost constant in the range @xmath41log(o / h)@xmath42 . it is important to remember that there are 243 points in this figure , therefore the lack of values on the @xmath33 relation over this particular range of metallicities can not be attributed to an incompleteness of the data ( compare also with fig . [ fig.5 ] where we have a similar number of normal h ii regions ) . in order to better understand what the excess of nitrogen abundance in sbngs means , we compare the distributions of oxygen and nitrogen abundances ( in percentages ) , both relative to their solar value . the distributions of oxygen abundances for the different samples , shown in fig . [ fig.7 ] , are similar . the samples cover almost uniformly a metallicity range between @xmath43 and @xmath44 dex . the fir - bright sbngs seem slightly more metal - rich than all the other galaxies in our sample . the distributions of nitrogen abundances , on the other hand , are significantly different in sbngs and in disk hii regions : a higher fraction of sbngs have solar abundances . the dispersion is relatively high for normal disk hii regions and their distributions peak at a value @xmath45 dex . this is shown in fig . [ fig.7]c and d. the dispersion of nitrogen abundances in the fir - bright sbngs is remarkably small . this constancy of n / o ( around log(n / o ) = -0.8 ) translates into the pseudo plateau seen in fig . [ fig.6 ] . next , we investigate the deviation @xmath46(n / o ) which is the difference between the observed nitrogen abundance and that predicted by the @xmath33 relation at the corresponding oxygen abundance . as shown in fig . [ fig.7]e and f , a high number of sbngs have a positive deviation from the @xmath33 relation . this confirms the peculiar structures observed in our data in fig . [ fig.6 ] . at first , it may seem surprising to find that what we define as `` normal '' hii regions have slightly lower n / o ratios than solar . but there is nothing which assures us that the solar value should be the standard in these objects . this result goes in the same sense as those based on studies of the metallicities of stars in the solar neighborhood : the most frequent metallicity ( estimated here by [ fe / h ] ) is around @xmath36 dex ( cayrel et al . 1997 ) and the sun seems to have a higher metallicity than the majority of stars ( eggen 1978 ) . the fact that the n / o distribution for sbngs peaks at the solar value implies a slight overabundance of nitrogen with respect to normal disk hii regions . but the range of n / o values found in the sbngs is comparable to that observed in the bulges of normal early - type spiral galaxies ( teh ; van zee et al . this is consistent with the recent discovery made by teh . these authors found that hii regions in early - type spirals have slightly higher n / o ratios than hii regions in late - type spirals . because the hii regions in late - type spirals are generally more numerous and luminous than in early - type spirals , samples of hii regions in normal galaxies are naturally biased towards those in late - type spirals , as can be easily verified for the mrs , vc&e and van zee et al . ( 1998 ) samples . the sbngs , on the other hand , are more numerous among early - type spirals ( see coziol et al . 1998a ) , explaining the observed higher abundance ratio in this sample . we conclude that our observations on the nitrogen abundance of sbngs are consistent with the chemical evolution of early - type spiral galaxies . it suggests that what we see could be the main production of nitrogen in the bulges of these galaxies . it is usually assumed that nitrogen is the product of the evolution of intermediate - mass ( @xmath44m@xmath47 m@xmath48m@xmath49 ) stars . consequently , the expected behavior of n / o as a function of o / h could be @xmath50 , @xmath33 or a mixture of both ( van den hoek & groenewegen 1997 ) . but a mixture of @xmath0 mode of production could also imply a contribution by massive stars . these two possibilities have very different consequences for the nature of sbngs . in the literature , we find some cases where a temporary contamination of nitrogen by massive stars is suspected ( pagel et al . 1992 ) . the contamination is supposed to come from massive wolf - rayet stars . but evidence in favor of a significant contribution of nitrogen by massive stars is still weak ( esteban & peimbert 1995 ) . in a recent paper , kobulnicky & skillman ( 1998 ) showed that , even in hii galaxies , which can be considered younger than sbngs , the nitrogen abundance seems to be correlated to the carbon abundance , which is inconsistent with a contribution by massive stars . for the more massive and evolved sbngs , a significant contribution of nitrogen by massive stars would imply major bursts of star formation , not much older than a few myrs . but , with the exception of the ultra - luminous infrared galaxies , observations of sbngs do not support such a scenario . their relatively low h@xmath5 equivalent widths , reddish @xmath51 colors , l(h@xmath5)/l(ir ) and l(b)/l(ir ) luminosity ratios are more consistent with low - level bursts occurring on a long or recurrent mode ( coziol 1996 ) . in general , sbngs also have relatively low rates of supernovae ( turatto et al . 1989 ; richmond et al . 1998 ; gonzlez delgado et al . 1999 ) , which is inconsistent with very young and massive bursts . another difficulty with the major burst scenario for sbngs is to identify the origin of such sudden and energetic events . the only mean we know of is interaction ( or merger ) between two massive spiral galaxies , like the ultra - luminous infrared galaxies for example . but most sbngs are relatively isolated and the origin of the bursts can not be blamed on a recent interaction with a nearby massive companion ( coziol et al . 1997a , 1997b ; contini et al . 1998 ) . the production of nitrogen by intermediate - mass stars is better documented and a more likely event . according to van den hoek & groenewegen ( 1997 ) , the bulk of nitrogen in galaxies is produced by agb stars during their thermal pulsing phase , when the stars lose most of their mass to the ism ( @xmath52 m@xmath49 for a 1 m@xmath49 star and @xmath53 m@xmath49 for a 6 m@xmath49 star ) . the low mass ( m @xmath54 m@xmath49 ) agbs contribute to he and c , while the high mass agbs contribute to he and n. the typical lifetime of a 4 m@xmath49 star is @xmath55 yrs , while that of a 8 m@xmath49 star is @xmath56 yrs ( charlot & bruzual 1991 ) . these time scales are much longer than the typical lifetimes of massive o and b stars . therefore , following this scenario , the bursts originate much further back in time , which solves the problem of the relative isolation . according to the model of charlot and bruzual ( 1991 ) , the early phase of a burst is dominated by massive stars of types o to b5 . the agb branch turns on after about 0.1 gyr . between 0.4 and 1.6 gyrs , the agb population accounts for @xmath57% of the bolometric luminosity , while massive stars of type b5 to a7 contribute @xmath58% . between 1.6 and 8 gyrs , the agb contribution drops to @xmath59% , and stars of type a7 to g0 dominate the main sequence . from this , we conclude that the main phase of production of nitrogen in a starburst should occur between 0.4 and 1.6 gyrs after the beginning of the burst ( this implies the evolution of stars from 8 to 3 m@xmath49 ) . now coziol ( 1996 ) found that the star formation in sbngs was nearly constant over the last 2 - 3 gyr period . thus , if we assume a median age of 1 gyr for the bursts , 2 or 3 bursts ( or more if the bursts have shorter durations ) could be sufficient to produce the number of co - evolved intermediate - mass stars needed for the abundance of nitrogen observed today . -0.25 cm -4.2 cm -9.5 cm in fig . [ fig.8 ] , we show schematically how a sequence of bursts could explain the production of nitrogen in the sbngs . our scenario is based on the analytical model presented in garnett ( 1990 ) . we assume that sbngs begin their chemical evolution with n / o and o / h ratios typical of hii galaxies . the first phase in their chemical evolution corresponds to an increase of o / h due to the evolution of massive stars . at the same time n / o decreases ( garnett 1990 ; olofsson 1995 ) . the second phase begins @xmath60 gyrs after the onset of the burst with the evolution of intermediate mass stars . because these stars produce mostly nitrogen , only n / o increases during this phase . the second burst begins with an increase in o / h , but with a lower amplitude than in the first burst . indeed , models of sequential bursts in galaxies usually assume that the successive bursts will have decreasing intensities ( gerola et al . 1980 ; krgel & tutukov 1993 ; marconi et al . 1994 : kppen et al . 1995 ) . the decrease in n / o during oxygen enrichment ( the slope of the vector ) is also smaller as it becomes more difficult to lower this ratio as the oxygen abundance increases ( garnett 1990 ) . again , about 0.4 gyr after the beginning of the second burst , the intermediate stars evolve and n / o increases , although not as much as in the first burst . the norms of the vectors in fig . [ fig.8 ] are related to the intensity of the bursts ; the stronger the burst , the higher the increase in o / h and n / o . the difference between the norm of the oxygen and nitrogen enrichment vectors should depend on the imf and on some other parameters related to the yields of the different stars . if we increase the number of bursts and assume that successive bursts are weaker and weaker , the sum of the vectors converges towards a line whose slope represents the mean increase of o / h and n / o in time , or in other words towards the @xmath33 relation . following this scenario , a constant star formation can be viewed as an infinite sum of very low - intensity bursts of star formation , explaining why normal disk hii regions follow the @xmath33 relation . -0.25 cm -4.2 cm -9.5 cm one prediction of this model is that the deviation of the observed n / o ratio from the @xmath33 relation ( @xmath46(n / o ) ) will decrease with the age of starburst galaxies , or , equivalently , with the increase in metallicity . in fig . [ fig.9]a , we see that this prediction is satisfied in the sbngs . we have verified that there are no relations between the deviation @xmath46(n / o ) and the ratios [ nii]/h@xmath5 and [ sii]/h@xmath5 , which rules out a possible explanation of this behavior in terms of an excitation effect ( indeed , following the excitation hypothesis the points with higher @xmath46(n / o ) would also have higher excess of emission in nitrogen and sulphur , which is not observed ) . the fact that the deviation in fig . [ fig.9]a is mostly positive could be explained by the different durations of the chemical evolutionary phases . because massive stars have very short lifetimes , the oxygen enrichment phase is almost instantaneous , and o / h rapidly reaches a maximum ( the tip of each horizontal vectors in fig . [ fig.8 ] ) . the lifetime of the stars producing nitrogen , on the other hand , spans a much larger range of values . the nitrogen enrichment probably increases rapidly at the beginning , but is prolongated over a longer period of time , as the lower - mass stars evolve . as a result , the top of each vertical vector in fig . [ fig.8 ] will always be much more populated , this phase representing a natural stable mode in the starburst s evolution . the above scenario for the formation of nitrogen in sbngs has some interesting implications that can be further tested . assuming that , in sbngs , carbon is mainly produced by intermediate mass stars , we should observe a deficiency of c with respect to n , because the agb stars that produce the carbon have longer main sequence lifetimes than those producing the nitrogen . then , by comparing the abundance of c with that of n we could determine the age of the sequence of bursts , while by comparing the abundance of n to that of s , we could determine the intensity and imf of the bursts . of course , none of this will be observed if carbon is mainly produced by massive stars ( maeder 1992 ) , its abundance increasing with metallicity , which is indeed observed in irregular galaxies ( garnett et al . 1995 ) and in the disks of _ normal _ spiral galaxies ( garnett et al . 1999 ) . we have shown in previous papers ( coziol et al . 1997b , 1998c ) , that the most plausible hypothesis for explaining the nearby sbngs is that they are the result of mergers of small - mass and gas - rich systems , probably similar to the nearby hii galaxies , which took place sometime in the recent past . the nearly constant star formation rate over a few gyr period ( coziol 1996 ) and now the time delay implied for the production of nitrogen by intermediate - mass stars both suggest merger ages of the order of 2 to 3 gyrs . on a cosmological time scale , this corresponds roughly to redshifts z @xmath61 to 0.3 ( using h@xmath62 km s@xmath63 mpc@xmath63 and assuming @xmath64 ) in the past . our hypothesis for explaining the nearby sbngs predicts a population of merging galaxies at intermediate redshifts . therefore , it is probably not a coincidence that we do find a substantial increase in the number of irregular blue star forming galaxies ( broadhurst et al . 1992 ; cowie et al . 1996 ; fioc & rocca volmerange 1997 ; lilly et al . 1998 ) at these redshifts . some of these galaxies have characteristics which are consistent with our interpretation for the sbngs , either in terms of star formation ( guzman et al . 1997 ; glazebrook et al . 1998 ; fioc & rocca volmerange 1997 ) or in terms of morphology ( colless et al . 1994 ; abraham et al . 1996 ; van den bergh et al . 1996 ; teplitz et al . 1998 ) . it is clear , on the other hand , that these pieces of evidence are not sufficient for identifying all the faint blue galaxies as the merging progenitors of the sbngs . there are still substantial problems to solve before understanding what the relation between these two phenomena could be . one of these problems , for instance , is the elapsed time between the two events . if some of the faint blue galaxies merged to form the nearby sbngs , they merged sometime between 2 or 3 gyrs ago . these time scales are much too long compared to the predictions of the best models of interacting - merging galaxies that we can find in the literature ( mihos & hernquist 1994 , 1996 for example ) . how could these galaxies sustain high rates of star formation over such long time scales ? we already proposed one solution , which is that the merging phase produces more than one burst of star formation . this could probably be achieved by introducing in the interacting - merging models different initial conditions , which would better suit the characteristics of proto - galaxies ( the merging galaxies are not massive spirals but small - mass , irregular , metal - poor and gas - rich galaxies ) , and by regulating the bursts of star formation with a mechanism like the feedback of supernovae ( gerola et al . 1980 ; krgel & tutukov 1993 ) or by repeating them in a sequence of mergers ( tinsley & larson 1979 ) . using the same empirical methods as for normal hii regions , we have estimated the n / o and o / h ratios of a sample of 243 hii regions observed in different sbngs . we have found clear evidence for an overabundance of nitrogen with respect to normal disk hii regions with comparable metallicities . the range of n / o values found in the sbngs is comparable to that observed in the bulges of early - type spiral galaxies . our observations are consistent with the discovery made by teh that the n / o ratios of h ii regions in the bulges of early - type spiral galaxies are higher than in late - type spirals . for the sbngs , this suggests that what we see could be the main production of nitrogen in their bulges . the most plausible hypothesis for the origin of nitrogen in sbngs is that it is the product of the evolution of intermediate - mass stars . these stars would have formed during a sequence of bursts extending over a few gyr period . following this scenario , the origin of the bursts is located 2 to 3 gyrs in the past . on a cosmological time scale these time intervals correspond to redshifts z @xmath61 to 0.3 , where one observes a substantial increase in the number of star forming galaxies whose characteristics are consistent with our hypothesis . this implies that some of the faint blue galaxies observed at intermediate redshifts may have merged to produce the sbngs observed today . the possible affiliation of the nearby sbngs with the excess of blue galaxies in the field at intermediate redshifts suggests that the sbngs are not a peculiar phase in the evolution of galaxies , but the result of a process which was much more common in the recent past of the universe . this process we tentatively identify with hierarchical formation of galaxies . we would like to thank dr . c. batalha who instructed us about the stellar abundance anomalies observed in the solar neighborhood and dr . r. r. de carvalho for very stimulating discussions on galaxies formation and cosmology . we thank the referee , dr . m. peimbert , for a critical reading and judicious suggestions . we are also thankful to dr . g. j. ferland for kindly providing the code cloudy and to the technical staff of observatoire de haute - provence for assistance at the telescope . carlos reyes acknowledges with thanks the receipt of research fellowships from the cnpq . coziol r. , contini t. , davoust e. , considre s. , 1998c , in abundance profiles : diagnostic tools for galaxy history , eds . d. friedli , m. edmunds , c. robert , and l. drissen , san francisco , asp conf . 147 , p. 219 storchi bergmann t.s . , wilson a.s . 1996 , in the interplay between massive star formation , the ism and galaxy evolution , d. kunth , b. guiderdoni , n. heydari malayeri , t.x . thuan , eds , ditions frontires , gif sur yvette ( france ) , p. 613 terlevich r.j . , tenorio - tagle g. , franco j. , boyle b.j . , 1991 , in the nature of compact objects in active galactic nuclei . proceedings of the 33rd herstmonceux conference . a. robinson , r. terlevich , cambridge university press , p. 209	we show that the excess of nitrogen emission observed in a large sample of starburst nucleus galaxies ( sbngs ) can only be explained at a given metallicity by an overabundance of nitrogen with respect to normal hii regions in the disks of late type spirals . the n / o ratios in the sbngs are comparable to the values found in the bulges of normal early type spirals , which suggests that what we observe could be the main production of nitrogen in the bulges of these galaxies . the variation of the n / o ratio as a function of metallicity in sbngs follows a @xmath0 relation , but the increase of nitrogen does not appear as a continuous process . in sbngs , nitrogen is probably produced by different populations of intermediate - mass stars , which were formed during past sequences of bursts of star formation . this assumption pushes the origin of the main bursts 23 gyrs back in the past . on a cosmological scale , this time interval corresponds to redshifts @xmath1 , where a significant increase of star formation activity occurred . the origin of the sbng phenomenon would thus have cosmological implications , it would be related to a more active phase of star formation in the universe sometime in its recent past .
the stellar mass is one of the most fundamental parameters characterising galaxies . this observable is driven by the light emitted in the rest - frame optical / near infrared ( nir ) by lower mass stars and it correlates with the dynamical mass up to @xmath20 ( e.g. , @xcite ) , suggesting that it can be a robust estimate of the cumulative content of matter in galaxies . stellar masses have been estimated for galaxies at redshifts as high as @xmath21 ( e.g. , @xcite ) . moreover , stellar mass estimates are readily available in the models of galaxy formation and evolution . for the above reasons , the stellar mass has been largely adopted in comparisons to the models . the stellar mass function ( smf ) , i.e. , the number density of galaxies per unit ( log ) stellar mass , provides a first census of a galaxy population and it is therefore one of the most basic observables that need to be reproduced by any successful model of galaxy formation . multi - wavelength photometric surveys like ultravista @xcite and zfourge @xcite have enabled smf measurements up to @xmath22 ( e.g. , @xcite , but see also @xcite for smf of massive galaxies up to @xmath17 ) , whereas _ hst _ surveys , like candels ( @xcite ) , goods ( @xcite ) and hudf ( @xcite ) , complemented by _ spitzer_/irac observations ( e.g. , spitzer / goods - pi dickinson ; @xcite ) , extended the study to the evolution of the smf up to @xmath17 ( e.g. , @xcite ) . estimates of stellar mass , however , critically depend on quantities like the initial mass function ( imf ) , the dust content , the metallicity and the star - formation history ( sfh ) of each galaxy ( see e.g. @xcite ) . indeed , different sets of sed models characterised by e.g. , a different treatment of the tp - agb phase , have been shown to potentially introduce systematics in stellar mass measurements as large as few decimal dex @xcite . similarly , different dust laws ( @xcite ) and star - formation histories ( sfhs ) can increase the systematics by up to @xmath23 dex ( @xcite ) . furthermore , recently , emission from nebular lines has been found to potentially bias stellar mass measurements ( @xcite ) . stellar masses of high redshift galaxies ( @xmath24 ) are particularly sensitive to the contamination by nebular lines , since high - z star - forming galaxies are likely to be characterised by emission lines with equivalent width in excess of @xmath41000 ( see e.g. , @xcite ) . nonetheless , attempts to correct for this contamination can lead to results differing by factors of a few ( e.g. , @xcite ) . alternatively , one could directly study the rest - frame optical / nir light emitted by the low - mass stars in galaxies , given its connection to stellar mass . measurements of the optical / nir luminosity have at least three major advantages over measurements of the stellar mass : 1 ) they are much less sensitive to assumptions about dust modeling ; 2 ) estimates of luminosity are robust quantities , since luminosity can be recovered directly from the observed flux , with marginal - to - null dependence on the best - fitting sed templates , and hence on e.g. , sfh or the imf ; and 3 ) a careful choice of the rest - frame band reduces the contamination by nebular emission , minimizing the requirement of corrections to the fluxes ( for instance , nebular emission could contribute up to 50% of the rest - frame @xmath25-band luminosity for galaxies at @xmath6 - @xcite ) . the wavelength range spanned by the rest - frame @xmath26 and @xmath27 bands is most sensitive to the lower mass stars . light in bands redder than these is potentially contaminated by emission from the dust torus of agns , while light in bluer bands retains information about the recent sfh . the availability of data up to @xmath28 m from _ spitzer_/irac has enabled the study of the evolution of the lf in rest - frame nir bands ( @xmath29 ) up to @xmath16 ( e.g. , @xcite ) . at higher redshifts , the choice of the rest - frame band which more closely correlates with the stellar mass must be a trade - off between including contamination from the recent star - formation and performing the luminosity measurements on observational data rather than relying on the extrapolation of sed templates . in this context , the @xmath3 band ( @xmath30 m ) emerges as a natural choice : it lays in the wavelength regime redder than the balmer/4000 break , it is free from contamination by strong nebular emission , and it can be probed up to @xmath6 thanks to the _ spitzer_/irac @xmath8m - band data . recently , @xcite have assembled the first full - depth irac mosaics over the goods and udf fields , combining irac observations from the igoods ( pi : oesch ) and iudf ( pi : labb ) programs with all the available archival data over the two fields ( goods , ers , s - candels , seds and udf2 ) . following the same procedure implemented by @xcite , full - depth irac mosaics have now been generated also for the goods - n and cosmos / ultravista fields . the goods - n mosaics double the area with the deepest irac imaging available , while the cosmos / ultravista data , shallower but covering a much larger field , are necessary to include the most massive galaxies at @xmath16 . the unique depth of irac @xmath31 m and @xmath8 m mosaics in the goods fields ( pid 194 ; pi dickinson ) - reaching @xmath32 ab ( @xmath33 , 20 aperture diameter ) allowed us to recover flux measurements with s / n@xmath34 in these two bands for galaxies to @xmath21 . in this work we leverage these characteristics to measure the evolution of the lf at @xmath19 in the rest - frame @xmath3 band , providing a complementary approach to the determination of the evolution of the smf at high redshift . furthermore , we will show that the smf can be recovered by applying a stellar mass - to - light ratio ( @xmath35 ) to the @xmath3-band lf . remarkably , a simple abundance matching reveals that the @xmath3-band lf can also trace the halo mass function ( hmf ) and its evolution over @xmath19 . our analysis is based on the photometric catalog of @xcite over the goods - n and goods - s fields . at @xmath16 we complement our sample with a 37-bands @xmath36-to-@xmath37 m photometric catalog based on the second release ( dr2 ) of the ultravista survey . the area covered by the dr2 data , @xmath38 degrees@xmath39 , is a factor of @xmath40 larger than the cumulative area from the two goods fields ( @xmath41arcmin@xmath39 ) , enabling the recovery of the bright end of the @xmath16 lf with higher statistical significance . this paper is organised as follows : in sect . 2 we describe the adopted datasets and sample selection criteria ; in sect . 3 we present our results . specifically , sect . 3.2 presents the stellar mass - to - light ratios from stacking , while our lf and smf measurements are presented in sect . 3.3 and 3.4 , respectively . we discuss our lfs measurements with respect to the halo mass function in sect . 4 , and conclude in sect . 5 . throughout this work we adopted a cosmological model with @xmath42 km / s / mpc , @xmath43 , @xmath44 . all magnitudes refer to the ab system . we assumed a @xcite imf unless otherwise noted . our lf measurements are based on a composite sample of galaxies at @xmath19 selected in the rest - frame optical ( @xmath3 band , @xmath30 m ) , with the bulk of our sample formed by lyman break galaxies ( lbgs ) from the candels / goods - n , candels / goods - s @xmath46 ers fields . the @xmath16 bin is complemented by a sample of galaxies from a catalog based on the ultravista dr2 mosaics . in the following sections we briefly describe these two parent catalogs and detail the criteria we adopted to assemble our final sample of galaxies . for this work we adopted the catalog assembled by @xcite over the candels / goods - n , candels / goods - s and ers fields . here we briefly summarise the main features referring the reader to sect . 2 and 3 of @xcite for full details . the catalog contains the photometry in the _ hst _ acs f435w , f606w , f775w and f850lp bands ( hereafter indicated by @xmath47 and @xmath48 ) , together with _ hst _ wfc3 f105w , f125w , f160w ( @xmath49 ) data from candels @xcite and wfc3 f140w band ( @xmath50 ) from the 3d - hst @xcite and aghast ( @xcite - http://mingus.as.arizona.edu/~bjw/aghast/ ) . the catalog takes also advantage of full - depth mosaics in the four _ spitzer _ irac bands . the @xmath7 m and @xmath51 m mosaics were assembled combining data from the igoods ( pi : oesch ) and iudf ( pi : labb ) programs to all the public archival data from either cryogenic or post - cryogenic programs over the goods - n and goods - s ( goods , ers , s - candels , seds and udf2 ) . for the @xmath31 m and @xmath8 m mosaics , instead , only observations from the goods cryogenic program are available ( pi : dickinson , pid : 194 ) . the mosaics were regenerated from the aors using the same procedure of @xcite . this procedure delivers the most accurate reconstruction of the psf at any position across each mosaic , enabling a more accurate flux measurement in the irac bands ( see below ) . the mosaics in the @xmath51 m , @xmath31 m and @xmath8 m bands are key for this work as they probe the rest - frame @xmath3 band . specifically , the @xmath51 m band matches the rest - frame @xmath3 band at @xmath16 , while the @xmath31 m and @xmath8 m bands are required for the rest - frame @xmath3 band at @xmath52 . figure [ fig : exptime_maps ] presents the exposure time maps in the irac @xmath51 m , @xmath31 m and @xmath8 m for the goods - n and goods - s fields . as a result of the combination of data from different programs the achieved depth across each field is highly inhomogeneous . this is particularly evident for the @xmath51 m band whose depth ranges between 50 hr and 180 hr ( corresponding to @xmath53 ab , respectively , for @xmath33 , 20 aperture diameter ) . the goods - n field is characterised by the deepest @xmath31 m and @xmath8 m data , reaching a depth of @xmath54 hr ( @xmath55 ab , @xmath33 , @xmath56 aperture ) . the object detection was performed on the @xmath57 image @xcite constructed from the @xmath49 band images . the detection mosaics have footprints of @xmath58 and @xmath59 arcmin@xmath39 , respectively for goods - n and goods - s , for a total of 264 arcmin@xmath39 . aperture photometry in the _ hst _ bands was performed in dual mode with sextractor @xcite on the mosaics matching the resolution of the @xmath60 image . fluxes were converted to total through the application of an aperture correction based on the @xcite scalable apertures and further corrected to take into account the flux losses of the scalable apertures compared to the point - spread function ( psf ) . photometry of the irac mosaics was performed using a proprietary deblending code ( @xcite ) . this code convolves the high - resolution _ hst _ mosaics with a kernel obtained from the highest signal - to - noise ( s / n ) irac psfs to construct a model of the irac image . for each object , @xmath56-diameter aperture photometry is performed on the image , previously cleaned from neighbours using the information from the model image . the aperture fluxes were then corrected to total using the hst template specific of each source convolved to match the _ spitzer _ irac psf . candidate lyman - break galaxies ( lbgs ) at @xmath61 and @xmath1 were selected among the @xmath62 and @xmath63 dropouts , respectively . for a complete list of criteria adopted to select each sample see table 2 of @xcite . the sample included 8031 lbgs . for the sample of @xmath16 galaxies , we also considered detections in the cosmos / ultravista field , whose larger field compared to goods - n / s allowed us to probe higher luminosities . the ultravista catalog used for this work is based on the _ ultradeep _ stripes of the second data release ( dr2 ) of the ultravista survey @xcite . this release is characterised by @xmath33 depth of @xmath64 , 25.1 , 24.8 , 24.8 ab ( @xmath56 aperture diameter ) in @xmath65 and @xmath27 , respectively ( @xmath66 mag deeper than dr1 ) and extends over an area of @xmath67 square degrees in 4 stripes over the cosmos field @xcite . the 37-bands catalog was constructed following the same procedure presented in @xcite for the dr1 . briefly , the detection was performed in the @xmath27 band ; 33-band far uv - to-@xmath27 aperture fluxes were measured with sextractor ( @xcite ) in dual mode on the mosaics matching the psf resolution of the @xmath26-band image . an aperture correction recovered from the kron ellipsoid was applied on a per - object basis ; total fluxes were finally computed by applying a further aperture correction obtained from the psf curve of growth . this new catalog also includes flux measurements in the subaru narrow bands nb711 , nb816 , the ultravista narrow band nb118 and the cfhtls @xmath68 , @xmath69 , @xmath70 , @xmath71 and @xmath3 , not available in the dr1 catalog of @xcite . the cosmos field benefits from several hundreds hours of integration time with _ spitzer _ irac . similarly to what was done for the goods - n / s fields , full - depth mosaics were constructed following the procedure of @xcite . specifically full depth 3.6@xmath72 m and @xmath51 m mosaics were reconstructed combining data from the s - cosmos @xcite , s - candels @xcite and splash ( pi : capak , @xcite ) . the resulting coverage map for the @xmath51 m band is shown in the lower panel of figure [ fig : exptime_maps ] . the depth ranges from @xmath73 to @xmath74 hours which correspond to @xmath75 ab ( @xmath33 in a @xmath56 aperture ) . observations in the @xmath31 m and @xmath8 m channels are only available from the s - cosmos _ spitzer _ cryogenic program . these data have a much shallower depth compared to the @xmath7 m and @xmath51 m bands , with an average limit of @xmath422.2 ab ( @xmath33 , 20 aperture ) . for this reason we only considered galaxies from the goods - n / s fields for the @xmath76 samples . fluxes in the four irac bands were measured using the template fitting procedure of @xcite , adopting the @xmath27 band as the high - resolution template image to deblend the irac photometry . photometric redshifts were computed using eazy @xcite on the 37 bands photometric catalog , complementing the standard eazy template set with a maximally red template sed , i.e. , an old ( 1.5 gyr ) and dusty ( @xmath77 mag ) sed template . we only considered objects whose fluxes were not contaminated by bright nearby stars , had extended morphology on the @xmath27 image , and less than five bands were excluded as the associated flux measurements were contaminated by ` nan ` values . galaxies in the @xmath16 redshift bin were selected among those with photometric redshift @xmath78 . the initial @xmath16 sample included 1208 objects . the photometric redshift selection allowed us to consider objects which could have been missed by a pure lbg selection . the large area offered by the ultravista dr2 footprint enabled the selection of bright / luminous sources whose surface density would be too low to be probed over the goods - n / s fields area . such luminous systems could be intrinsically redder than normal lbgs either because they are more ( massive ) evolved systems and/or they contain a higher fraction of dust . on the other side , the lbg selection at fainter luminosities from the goods - n / s samples is expected to suffer only limited selection bias against intrinsic red sources as in this range of luminosities galaxies are mostly blue star - forming systems , with low content of dust . band of our composite sample of galaxies at @xmath79 and @xmath1 , as labeled in the figure . for the @xmath16 sample , the histograms for the goods - n / s and ultravista samples are also presented separately , showing the complementarity in @xmath9 of the two datasets . [ fig : absm_histogram],width=377 ] the first step consisted in applying a cut in the @xmath27 flux of the galaxies from the ultravista sample , in order to control the detection completeness . the dr2 data is @xmath80 mag deeper than dr1 . therefore , we set the threshold to @xmath81 mag , corresponding to the 90% completeness in detection of point - sources @xcite . instead , we did not apply any cut in the detection band of the goods - n / s sample , as the method adopted to estimate the co - moving volumes already takes into account the incompleteness from the detection stage . successively , we excluded from our sample those galaxies with poor flux measurements in those irac bands used to compute the rest - frame @xmath3 luminosity . the variation in depth across each irac mosaic prevented us from applying a single value of flux threshold at this stage . instead , we applied a cut in s / n to the flux in the irac band closest to the rest - frame @xmath3 band ( i.e. , irac @xmath51 m at @xmath16 and irac @xmath31 m and @xmath8 m at @xmath52 ) . considering the gap in the photometric depth probed by the @xmath51 m mosaics compared to that reached by the @xmath82 m data , we opted for applying a distinct s / n cut depending on the considered redshift bin . the sample at @xmath16 was selected by applying the cut of @xmath83 to the @xmath51 m flux ; the samples at @xmath84 were assembled by considering the cumulative flux in the @xmath31 m and @xmath8 m bands as the inverse - variance weighted sum of the flux in these two bands . we then applied a cut to the corresponding s / n such that : @xmath85 where s@xmath86 is the flux measurement in band @xmath87 and @xmath88 is the weight defined as @xmath89 , with @xmath90 the corresponding flux uncertainty . the application of the s / n cut reduced the number of galaxies to 2644 ( 2040/604 for goods / ultravista , respectively ) , 96 , 17 and 4 at @xmath16 , @xmath91 , @xmath92 and @xmath17 , respectively . we further cleaned our sample , excluding those objects satisfying any of the following conditions : 1 ) the contribution to the @xmath31 m and @xmath8 m flux from neighbouring objects is excessively high ; 2 ) the source morphology is very uncertain or confused making irac photometry undetermined ; 3 ) the source is detected at x - rays wavelengths , suggesting it is a lower redshift agn ; 4 ) the source is at higher redshift , but its sed is dominated by agn light ; 5 ) lbgs with a likely @xmath93 solution from photometric redshift analysis . in appendix [ app : sel ] , we detail our application of these additional criteria in cleaning our sample . our final sample consists of 2098 galaxies at @xmath16 ( 1680 from the lbg sample and 418 from the ultravista sample ) , 72 at @xmath91 , 10 at @xmath92 and 2 objects at @xmath17 . the distribution of the absolute magnitudes in the @xmath3 band for the sample is presented in figure [ fig : absm_histogram ] , for the four different redshift bins . it is noteworthy how the goods - n / s and ultravista samples complement each other at @xmath16 , allowing to fully exploit these data with little redundancy . m ( top panel ) and @xmath31m+@xmath8 m ( bottom panel ) bands . the pink curve presents the fraction of objects excluded because their flux measurements were highly contaminated by neighbours . the pink shaded area presents the associated @xmath94 poisson uncertainties . the blue curve and shaded area present the completeness fraction from the s / n cut in the flux of the corresponding band and associated poisson uncertainty , while the black curve and grey shaded area show the combined effect of s / n threshold and contamination cleaning . [ fig : completeness],width=340 ] we implemented a monte carlo simulation based on real data to estimate the effects that our selection criteria in s / n and contamination polishing have on the sample of galaxies used in this work . for this simulation , we did not consider the effects of selection in the detection band because the ultravista sample is 90% or more complete in @xmath27 by construction , while the effects of detection completeness of the goods - n / s sample have been taken into account when estimating the co - moving volumes adopted for the lf measurements . at first we defined a grid in apparent magnitude of width 0.20 mag . given the small sizes of the galaxies compared to the irac psf , for each magnitude value in the grid , we injected 100 point sources randomly distributed across a region of uniform depth in the @xmath51 m , @xmath31 m and @xmath8 m mosaics of the goods - n field . we chose the goods - n as this field is characterised by the deeper _ spitzer_/irac @xmath51 m , @xmath31 m and @xmath8 m band data among the fields considered for this work . successively , we replicated the flux measurement using the same procedure adopted for the actual photometry . the completeness fraction in each magnitude bin was computed by comparing the number of objects satisfying our selection criteria ( section [ sect : sample_assembly ] ) to the number of objects initially injected into the simulation . for the completeness of the @xmath16 sample , the above process was applied to the @xmath51 m mosaic only . for the completeness of the samples at @xmath76 , the point sources were added at matching positions in the @xmath31 m and @xmath8 m mosaics . the selection on the s / n and contamination was then recovered applying the corresponding criteria and assuming the sed to be flat in @xmath95 in the observed @xmath96 m region . this is a reasonable approximation since , as we show in section [ sect : ml_ratios ] , the median seds do not substantially deviate from a flat @xmath95 sed in the wavelength range covered by irac observations . the whole process was repeated @xmath97 in each band in order to increase its statistical significance . the global completeness ( i.e. , the cumulative effects of s / n and contamination selection ) at the different depths of the irac mosaics was obtained rescaling the completeness in s / n selection to match the depth of the relevant region . the results from our completeness simulation for the @xmath51 m and the @xmath31m+@xmath8 m samples are presented in figure [ fig : completeness ] . first we discuss the recovery of the contamination fraction ; successively we consider these results in the budget of the gobal completeness estimates . the fraction of objects in the @xmath51 m band contaminated by neighbours is negligible for objects brighter than @xmath98 ab and increases exponentially up to @xmath99 ab , where it starts to flatten out . a similar behaviour is observed for the @xmath31m+@xmath8 m simulation , although shifted at brighter magnitudes , due to the shallower depth of the @xmath31 m and @xmath8 m compared to the @xmath51 m . the flattening at the faint end is caused by a strong incompleteness in the data at such faint magnitudes and likely does not reflect the true behaviour . in what follows and in our analysis we do not consider the completeness for magnitudes fainter than those corresponding to the onset of the flattening , i.e. , @xmath99 ab and @xmath100 ab for the @xmath51 m and @xmath31m+@xmath8 m data respectively . as it could intuitively be expected , the bright end of the global completeness curve is dominated by the ( small ) fraction of purged objects . this effect becomes less and less pronounced at fainter magnitudes , corresponding to lower s / n , where the effective selection is driven by the s / n itself . the random noise from the background can positively combine at the location of a given source , introducing an increase in the measured flux ( _ flux boosting _ , @xcite ) . the amount of this boost is inversely correlated to the s / n of the source . the flux for sources with very high s / n will mostly be the result of the photons emitted by the source itself , with reduced contribution from the background ; on the other hand , for sources with low s / n , the background level can be typically just few factors smaller than the intrinsic signal from the source , making it sensitive to ( positive ) fluctuations of the background . furthermore , sources do not uniformly distribute with flux , but rather follow a @xmath4 power - law , with fainter sources more numerous than brighter ones . therefore it is intrinsically more probable that fainter sources scatter to brighter fluxes than the reverse , giving origin to a net flux bias . a second potential source of flux boosting comes from confusion noise : faint sources at apparent positions close to a brighter one are more likely to be blended into the brighter source , increasing the flux , and decreasing the number of fainter objects . this effect is larger for flux measurement in those bands with wide psf , like _ spitzer_/irac . however , in our case , the photometry in the irac bands was performed adopting a higher resolution morphological prior from _ hst _ mosaics ( see sect . [ sect : data_sets ] ) . furthermore , we applied a selection in flux contamination ( see sect . [ sect : sample_assembly ] ) . since these two factors drastically limit the potential contribution of confusion noise to the irac fluxes in our sample , we do not further consider its effects to the flux boosting budget . for each source , we estimated the flux bias as the ratio between the expected intrinsic flux @xmath101 and the measured flux @xmath102 . since no direct measurement is possible , the intrinsic flux was recovered as the average flux obtained from an estimate of its probability distribution . this was constructed considering two distinct contributions : 1 ) the probability @xmath103 that the observed flux @xmath102 is drawn from the distribution of intrinsic flux @xmath101 , given the noise @xmath104 , and 2 ) the frequency @xmath105 of occurrence of the intrinsic flux @xmath101 . assuming each probability is normalized to 1 , the final probability distribution would then be @xmath106 . assuming a gaussian noise , @xmath103 can be written as : @xmath107 } , \label{eq : gaussian}\ ] ] normalized to a total probability of 1 . the frequency associated to the intrinsic flux can be recovered from the ( intrinsic ) differential number count of sources , @xmath108 . this can usually be described by a power - law form with negative index , thus divergent for @xmath109 , which prevents it from being normalized , the product between @xmath110 and @xmath111 does not diverge for @xmath109 . however , this is not anymore the case for lower s / n values , where the non - negligible probability of the low - flux tail from the gaussian distribution makes the divergent power - law dominate over the gaussian . ] ( see e.g. , @xcite , who also discuss possible reasons for why the divergence at @xmath109 is likely non - physical ) . we therefore followed the formalism of @xcite , who introduced as further constraint the poissonian probability that no sources brighter than @xmath101 exist at the same location of the observed object . the expression for @xmath105 then becomes : @xmath112 where @xmath113 corresponds to the area occupied by the source . in the left panel of figure [ fig : flux_boost ] we show examples of reconstructed @xmath114 , for the cases of s / n=6 and s / n=3.8 , where it is evident the increasing contribution of the faint sources population to the expected intrinsic flux as the s / n decreases . the right panel of figure [ fig : flux_boost ] shows the expected flux boost as a function of s / n . for s / n@xmath115 , the flux boost is roughly the same amount as the flux uncertainty . however , for lower s / n the estimated flux boost increases abruptly . for s / n@xmath116 the expected flux boost is @xmath117 mag , meaning that the recovery of the intrinsic flux for such low s / n data is highly uncertain . the s / n in the _ hst _ bands for the galaxies in our sample is @xmath118 . at @xmath16 the s / n in the @xmath51 m band , adopted for the selection of the @xmath16 sample , is @xmath119 by construction ; the s / n in the @xmath7 m band is @xmath119 as well , consistent with the nearly flat seds in that wavelength range . at @xmath76 , the selection was performed in a combination of @xmath31 m and @xmath8 m fluxes , adopting a s / n@xmath120 threshold . figure [ fig : flux_boost ] shows that the expected flux boost for s / n@xmath121 is @xmath122 mag . however , for lower s / n , typical of the selection of samples at @xmath76 , the correction can be as high as @xmath123 mag . we therefore applied the above correction to the fluxes in the @xmath31 m and @xmath8 m of the @xmath76 samples . the average flux boost was @xmath124 mag and @xmath125 mag in the irac @xmath31 m and @xmath8 m bands , respectively . in appendix [ app : seds ] we present the seds of the 12 most luminous galaxies in the @xmath91 sample together with the seds of the @xmath92 and @xmath17 samples , before and after applying the flux boost correction . in the last few years a number of works have studied the relation between the rest - frame uv luminosity and the stellar mass of high - redshift galaxies ( e.g. , @xcite ) . indeed , a relation between the stellar mass and the uv luminosity is to be expected considering a continuous star formation . deviations from such a relation would then provide information on the age and metallicity of the stellar population and on the dust content of the considered galaxies . the emerging picture is that at @xmath16 and for stellar masses @xmath126 , the stellar mass increases monotonically with increasing uv luminosity ; however , at stellar masses higher than @xmath127 the trend becomes more uncertain : @xcite using a sample of @xmath27-based photometric redshift selected galaxies found indication of a turnover of the uv luminosity , with the more massive galaxies ( @xmath128 ) spanning a wide range in uv luminosities ( see also @xcite ) ; @xcite , instead , using a lbg sample , found a linear relation between uv luminosities and stellar masses up to @xmath129 . considering the different criteria adopted by the two teams for the assembly of their samples , selection effects might be the main reason for the observed tension . a monotonic relation between the uv luminosity and the stellar mass has also been found at @xmath91 and @xmath92 for @xmath126 ( e.g. , @xcite ) , with approximately the same slope and dispersion , but with an evolving normalisation factor ( but see e.g. , figure 5 of @xcite for further hints on the existence of massive galaxies with faint uv luminosities ) . figure [ fig : uvopt ] presents the absolute magnitude in the rest - frame @xmath3 band ( @xmath9 ) as a function of the absolute magnitude in the uv ( @xmath130 ) , for our sample in the four redshift bins ( @xmath61 and @xmath1 ) , while in figure [ fig : median_uvopt ] we present the binned median in the @xmath131 plane for the @xmath16 and @xmath91 samples . the @xmath16 sample shows a clear correlation between the luminosities in the rest - frame uv and @xmath3 bands for @xmath132 mag , which can be described by the following best - fitting linear relation : @xmath133 the above best - fit is marked by the magenta line in figure [ fig : median_uvopt ] , where we also indicate the @xmath134 limits corresponding to our @xmath31m@xmath135 m selection . at @xmath16 and @xmath91 the depth of the irac data allows us to not only probe the bright end , where the relation between @xmath136 and @xmath9 breaks , but to also explore the regime of the linear correlation expressed by eq . [ eq : uvopt ] . slopes of 0.4 - 0.5 in the @xmath137 plane ( with nominal @xmath138 uncertainties of @xmath139 ) have been reported by e.g. , @xcite and @xcite . assuming a constant @xmath13 ratio ( see section [ sect : ml_ratios ] ) , our measurements correspond to a slope of 0.5 in the @xmath137 plane , consistent with previous measurements . assuming that the sfr mostly comes from the uv light and that @xmath9 is a good proxy for stellar mass measurements , we can also compare the slope we derived for the @xmath140 relation to that of the log(sfr)-@xmath141 from the literature . indeed , the observed uv slopes of @xmath142 galaxies in our sample are @xmath143 ( see also @xcite ) , consistent with star forming galaxies and little - to - no dust extinction . our measurement is perfectly consistent with the log(sfr)-@xmath141 slope of @xmath144 recently measured by e.g. , @xcite for @xmath145 star forming galaxies with @xmath146 , and it has been shown to evolve little over the redshift range @xmath147 . absolute magnitudes versus the uv absolute magnitudes for our composite sample , color - coded according to the considered redshift bin . the @xmath16 data are presented as a density plot , with denser regions identified by a darker color , while the points for the @xmath76 samples are shown individually . the uv-@xmath3 relation shows a turnover for @xmath148 . [ fig : uvopt],width=340 ] vs. @xmath9 relation . the blue points mark the median at @xmath16 , while the green points mark the median at @xmath91 . the lower and upper blue ( green ) curves represent the 25 and 75 percentiles , respectively , of the points at @xmath16 ( @xmath91 ) . the vertical green , yellow and red shaded regions indicate the limiting magnitude corresponding to our s / n cuts at @xmath91 , @xmath92 and @xmath17 , respectively . the magenta line represents the best - fit relation for the @xmath16 sample in the range @xmath149 mag . [ fig : median_uvopt],width=359 ] for absolute magnitudes brighter than @xmath150 mag the linear relation expressed by eq . [ eq : uvopt ] breaks , as we observe the beginning of a turnover in the absolute uv-@xmath3 magnitude relation . remarkably , this behaviour is visible also for the @xmath91 sample , which is entirely based on lbg selection . this fact has important consequences for e.g. , stellar mass function measurements : samples of galaxies selected at fixed rest - frame uv luminosity are potentially characterised by a wide range of stellar mass . the absolute uv magnitude of galaxies with @xmath151 mag spans the full range of values observed for @xmath148 mag . however , the bulk of values aggregates around the @xmath152\sim[-21.4 , -23.5]$ ] mag region and it is characterised by a large dispersion in @xmath136 ( @xmath153 mag ) . this result is qualitatively consistent with what found by @xcite , assuming a correlation between the absolute magnitude @xmath9 and the stellar mass . most of the galaxies with @xmath154 come from the photometric redshift sample selected from the ultravista catalog . as we will present in sect . [ sect : ml_ratios ] , our measurements of the mass - to - light ratios from stacking analysis show that galaxies with @xmath155 statistically have stellar masses @xmath156 . the above result then underlines the bias that lbg selections may introduce against massive systems . at @xmath91 our data allows us to inspect the relation only for @xmath9 fainter than @xmath157 and @xmath158 fainter than @xmath159 . in this range of luminosities , our @xmath91 measurements are roughly consistent with the @xmath16 measurements in the same range of luminosities . the measurements for the @xmath92 and @xmath17 samples are still consistent with the trends observed at @xmath16 , although the low number of objects does not allow us to derive any statistically significant conclusion . so far determinations of the stellar mass - to - light ratios ( @xmath35 ) for galaxies at @xmath160 have involved the @xmath161 ratio . this quantity is fundamental for our understanding of galaxy formation and evolution as it combines information on the recent ( through the uv luminosity ) and on the integrated ( through the stellar mass ) sfh ( e.g @xcite ) . nonetheless , the @xmath161 has been used to recover the stellar mass and smf of high redshift galaxies with alternating success ( see e.g. , @xcite ) . in this section , instead , we explore for the first time the @xmath13 properties of galaxies at @xmath24 . the rest - frame @xmath3 luminosity is more sensitive to the stellar mass compared to the uv luminosity for two reasons . while the uv light is emitted by massive , short - lived stars and thus traces the sfh in the past few hundred myr , the luminosity in the rest - frame optical region mostly originates from lower mass , longer living stars , which constitute most of the stellar mass of galaxies . furthermore , it is less sensitive to the dust extinction , and hence to the uncertainties in its determination , compared to the uv : for a @xcite extinction curve , an @xmath162 mag gives @xmath163 mag compared to @xmath164 mag . since here we are interested more on average trends in the @xmath13 ratios rather studying it for specific galaxies , we performed our analysis using the median stacked seds constructed from our composite sample . due to the different photometric bands in the catalogs , we performed the stacking of sources separately for sources in the goods - n / s and ultravista samples . the stacked seds were constructed as follows ( see also @xcite ) . at each redshift interval we divided the galaxies into sub - samples according to their @xmath9 . the different depths reached by the @xmath51 m and @xmath31m@xmath135 m samples resulted in different number of subsamples across the redshift bins . under the working assumption of limited variation in both redshift and sed shape in each bin of @xmath9 , and for each _ hst _ band we took the median of the individual flux measurements . our assumption is also supported by the fact that the seds from stacking are generally characterized by a flat @xmath95 continuum at both rest - frame uv and optical regimes . uncertainties on the median were computed from bootstrap techniques , drawing with replacement the same number of flux measurements as the number of galaxies in the considered absolute magnitude bin . before median - combining , the fluxes were perturbed according to their associated uncertainty . the process was repeated 1000 times and the standard deviation of the median values was taken as the final uncertainty . for the irac bands , median stacking was performed on the mosaic cutouts centered at the position of each source , previously cleaned from neighbours . photometry was performed on the median of the images in apertures of @xmath165 diameter . total fluxes were then recovered through the psf growth curve . uncertainties were computed applying to the image cutouts the same bootstrap technique adopted for the median stacking of the fluxes , as described above . in randomly drawing the image cutouts , we preserved the total exposure time . photometric redshifts and @xmath3-band luminosities were obtained from eazy ( @xcite ) on the stacked seds ; stellar masses were computed running fast ( @xcite ) adopting the @xcite template seds , a @xcite imf , solar metallicity and a delayed - exponential sfh . the bands potentially contaminated by nebular emission were excluded from the fit . since we performed the stacking in each band individually , assuming the same redshift for all sources , the flux in those bands close to the lyman and the balmer breaks potentially suffers from high scatter , introduced by the range of redshifts of the galaxies in each sub - sample , and depending on whether the break enters or not the band . fluxes in these bands were therefore excluded from the fit with fast . specifically , for the @xmath16 lbg stacks , we excluded the @xmath166 and @xmath60 bands ; for the @xmath16 ultravista stacks , we excluded the @xmath167 @xmath168 @xmath169 bands ; for the @xmath91 stacks , we excluded the @xmath170 and the @xmath171 bands ; for the @xmath92 stack , we excluded the @xmath171 band while for the @xmath17 stack we excluded the @xmath172 band . the stacked seds together with the best - fit templates from fast are presented in figure [ fig : stack ] . the photometric redshifts measured from the stacked sed are all consistent with the values of the corresponding redshift bin ; the difference of the photometric redshifts of the median stacked and the median of the photometric redshifts of the individual sources in each subsample is @xmath173 , i.e. within the uncertainties expected for photometric redshifts . the stacked seds at @xmath16 and @xmath91 are characterised by a trend in the uv slope , with bluer slopes for low - luminosity galaxies , particularly evident for the @xmath16 stacks , and qualitatively consistent with the results of e.g. , @xcite and @xcite . furthermore , the seds in the rest - frame uv wavelength region of the 4 - 5 brightest @xmath16 stacks do not differ too much one from each other , while they differ substantially at wavelengths redder than the balmer/4000 break . this trend is reproduced in the inset of the @xmath16 panel of figure [ fig : stack ] , where we plot the magnitude in the observed @xmath51 m band , @xmath174 ( @xmath4 rest @xmath3 ) , versus the magnitude in the observed @xmath175 band , @xmath176 ( @xmath4 rest uv ) . it is clear that the stacks from the lbg selection present a roughly linear relation between the magnitudes in the two bands . this trend is interrupted at the brightest @xmath176 , where @xmath176 does not follow anymore the brightening of @xmath174 . ) from our stacking analysis as a function of absolute magnitude in the @xmath3 band ( @xmath9 ) . the color codes in the legend identify the four redshift bins considered in this work . the assumed constant and best fit log - linear relations to the @xmath16 points are displayed by the grey lines and are expressed by the grey labels . [ fig : ml],width=340 ] recently , @xcite presented stacked seds at @xmath16 in bins of @xmath177-band absolute magnitudes from a sample of galaxies based on candels / goods - s , hudf and hudf09 - 2 . the sample benefits from deep irac @xmath7 m and @xmath51 m imaging from the irac ultra deep field ( iudf ) program . the stacked seds ( see e.g. , their fig . 2 ) show a clear trend of redder colors with increasing rest - frame @xmath177-band luminosity , in particular for @xmath178 . our stacked seds are in qualitative agreement with those of @xcite , confirming the observed trend with luminosity . furthermore , thanks to the wide area offered by ultravista which provides coverage for even brighter sources , we are able to extend the trend to even more luminous galaxies . in figure [ fig : ml ] we present the @xmath13 measured from the stacked seds as a function of @xmath3 absolute magnitude . total uncertainties were obtained by propagating the 68% confidence intervals in stellar mass generated by fast and the uncertainties in luminosity , taken as the flux uncertainties from stacking . at @xmath16 the @xmath13 ratio is consistent with being constant for @xmath9 fainter than @xmath179 mag . we find : @xmath180 for @xmath148 mag there is indication of @xmath13 increasing with the luminosity , although the error bars are large . the best - fit seds of the most luminous stacks have a nearly constant age ( @xmath181 yr ) , and show an @xmath182 slightly increasing with stellar mass ( from 1.0 to 1.2 mag ) . a linear fit of the @xmath183 values for @xmath148 mag resulted in the following relation : @xmath184 our linear relation recovers the constant value of @xmath185 at @xmath186 . however , the uncertainties on the fit parameters make the above relation also consistent with a constant value . the @xmath13 ratios for the @xmath91 and @xmath92 are consistent with @xmath13 measurements of the @xmath16 , @xmath187 ab stacks . the @xmath13 for the @xmath17 bin is consistent with the average @xmath13 only at @xmath188 level . we note however , that the @xmath17 sample only includes two sources , therefore reducing the statistical significance of the observed disagreement . the above results are consistent with what already observed in figure [ fig : uvopt ] . in sect . [ sect : uvopt ] we showed that galaxies more luminous than @xmath186 mag form a cloud in the rest - frame uv-@xmath3 plane around @xmath189 mag . from our stacking analysis , the average apparent magnitude at @xmath190 ( i.e. , the rest - frame uv@xmath191 ) for the stacked seds with @xmath192<23.5 $ ] ab is @xmath193 ab , which at @xmath16 corresponds to an absolute magnitude @xmath194 . according to the above relation , the stellar mass corresponding to @xmath195 mag is @xmath196 . this behaviour warns about the potential biases that can occur when adopting the uv luminosity and @xmath197 in the measurement of stellar masses , in particular for massive galaxies . a constant @xmath35 is equivalent to a slope of @xmath198 in the log(stellar mass ) - absolute magnitude plane . our result at @xmath16 , obtained for galaxies with @xmath199 mag , is consistent with the @xmath200 slopes found in the stellar mass - @xmath201 plane ( see e.g. , @xcite ) . steeper slopes , as those found by @xcite , require the @xmath202 to decrease for fainter galaxies or increase for brighter galaxies or a combination of both effects . the origin for this is still unclear as it could be a mix between selection effects ( see e.g. @xcite ) and nebular emission contamination which could boost the stellar masses of the more luminous galaxies ( e.g. , @xcite ) . cccdr @xmath177 & @xmath9 & @xmath203 & 2c@xmath204 & # + bin & ( mag ) & ( mag ) & 2c(@xmath205 mpc@xmath206 mag@xmath207 ) & gal . + @xmath16 & @xmath208 & 0.25 & @xmath209 & 1 + & @xmath210 & 0.25 & @xmath211 & 2 + & @xmath212 & 0.25 & @xmath213 & 5 + & @xmath214 & 0.25 & @xmath215 & 16 + & @xmath216 & 0.25 & @xmath217 & 29 + & @xmath218 & 0.25 & @xmath219 & 39 + & @xmath220 & 0.25 & @xmath221 & 55 + & @xmath222 & 0.25 & @xmath223 & 78 + & @xmath224 & 0.25 & @xmath225 & 100 + & @xmath226 & 0.25 & @xmath227 & 86 + & @xmath228 & 0.25 & @xmath229 & 64 + & @xmath230 & 0.25 & @xmath231 & 72 + & @xmath232 & 0.25 & @xmath233 & 64 + & @xmath234 & 0.25 & @xmath235 & 87 + & @xmath236 & 0.25 & @xmath237 & 123 + & @xmath238 & 0.25 & @xmath239 & 128 + & @xmath240 & 0.25 & @xmath241 & 186 + & @xmath242 & 0.25 & @xmath243 & 194 + & @xmath244 & 0.25 & @xmath245 & 196 + & @xmath246 & 0.25 & @xmath247 & 207 + & @xmath248 & 0.25 & @xmath249 & 193 + & @xmath250 & 0.25 & @xmath251 & 119 + & @xmath252 & 0.25 & @xmath253 & 55 + & & & . & + @xmath91 & @xmath220 & 0.50 & @xmath254 & 2 + & @xmath224 & 0.50 & @xmath255 & 9 + & @xmath228 & 0.50 & @xmath256 & 16 + & @xmath232 & 0.50 & @xmath257 & 20 + & @xmath236 & 0.50 & @xmath258 & 17 + & @xmath240 & 0.50 & @xmath259 & 8 + & & & . & + @xmath92 & @xmath260 & 0.80 & @xmath261 & 1 + & @xmath230 & 1.00 & @xmath262 & 4 + & @xmath263 & 0.70 & @xmath264 & 5 + & & & . & + @xmath17 & @xmath232 & 0.50 & @xmath265 & 2 + the lfs were measured adopting the @xmath266 estimator @xcite . although this method is intrinsically sensitive to local overdensities of galaxies , at @xmath160 the clustering is expected to be negligible . on the other hand , the @xmath266 method directly provides the normalisation of the lf . furthermore , and most importantly , the _ coherent analysis _ extension developed by @xcite is key to this work . as we showed in sect . [ sect : data_sets ] , our composite sample is based on a dual - band flux selection , corresponding to a double flux threshold . the detection process introduces the first flux cut in the corresponding band ( @xmath27 or @xmath57 image built from the hst nir bands , for the ultravista and goods - n / s sample , respectively ) . the s / n cut on the flux in the irac band closest to the rest - frame @xmath3 is responsible for the second flux threshold in the relevant irac band . for each galaxy in the sample , each flux threshold generates an upper limit to the redshift the specific galaxy can have and still be included in the sample . these different upper limits in redshift correspond to different comoving volumes for each object which could potentially enter the @xmath267 computation . the coherent approach allowed us to take this double selection into consideration in a consistent way : the upper limit in redshift , used to compute the comoving volume , was taken to be the smaller one among the two redshift upper limits computed based on the threshold in the corresponding selection band . furthermore , as we showed in figure [ fig : exptime_maps ] , the depth of the irac mosaics is highly inhomogeneous . therefore , for the computation of the comoving volumes in each field , we divided the irac footprint into a number of sub - fields , such that each sub - field was characterised by nearly homogeneous depths in both the detection image from the combination of different filters , we considered the wfc3/@xmath60 the relevant band for the depth of the detection in the goods - n / s fields . ] and in the relevant irac band . again , the @xcite prescription allowed us to analyse the different sub - samples coherently . comoving volumes were computed differently depending on the field and on the band driving the selection . for the galaxies in the goods - n / s fields , we used the comoving volumes computed by @xcite . these volumes were estimated using an extensive monte carlo simulation based on real data . sources were added to the different mosaics and recovered following the same procedure applied for the assembly of the lbg sample . such volume estimates natively take into account the selection effects at the detection stage , correct for flux - boosting effects and contamination by lower redshift interlopers and brown dwarfs . the volumes @xmath268 for those objects @xmath175 in the goods - n / s fields whose redshift upper limit @xmath269 was driven by the irac s / n threshold ( @xmath270 ) were rescaled by the ratio between the volume associated to the redshift upper limit from the irac band @xmath271 and the redshift upper limit in the @xmath57 image @xmath272 : @xmath273 for the ultravista sample , the volumes were computed directly from the limits in redshift corresponding to the flux limits in the @xmath27 and @xmath51 m bands . we computed the lf in four redshift bins centred at @xmath16 , @xmath91 , @xmath92 and @xmath17 . although the irac data potentially allowed us to consider galaxies at @xmath6 , we did not find any candidate with reliable flux measurement in the irac @xmath31 m and @xmath8 m . uncertainties on the lf measurements were derived by combining in quadrature the poisson noise in the approximation of @xcite to an estimate of cosmic variance from the recipe of @xcite . the average cosmic variance value obtained for the @xmath16 ultravista sample was @xmath274 ; the average cosmic variance estimates for the goods - n / s sample were @xmath275 and @xmath276 , respectively for the @xmath79 and @xmath17 redshift bins . the high values of the cosmic variance registered for all redshifts and luminosities are the dominant source of stochastic uncertainties in our lf measurements . our lf measurements are presented in figure [ fig : lf ] and in table [ tab : lf ] . the absolute magnitude range of the @xmath16 lf spans @xmath277 magnitudes , @xmath278 more than the magnitude range of the @xmath76 lfs . the larger absolute magnitude range available at @xmath16 is the result of a number of distinct factors . first , the increased depth in the @xmath51 m band from the combination of _ spitzer_/irac cryogenic and post - cryogenic epochs enables to reach fainter absolute magnitudes than the cryogenic only @xmath279 m data at @xmath280 . second , the smaller psf size of the @xmath51 m data compared to the @xmath31 m and @xmath8 m bands allows to reach fainter fluxes for the same exposure time and detector efficiency . third , the availability of cosmos / ultravista data over an area @xmath281 larger than the goods - n / s footprint allowed us to recover the exponential decline of the bright end of the @xmath16 lf , otherwise inaccessible by the small footprint of the goods - n / s mosaics . in order to verify the consistency of the @xmath16 lf with respect to the goods - n / s and ultravista data , we also computed the lf separately on each one of these two datasets . the resulting lfs are marked in figure [ fig : lf ] by the open squares and show a good agreement with the lf from the composite sample . the large uncertainties associated to the number density measurements at @xmath84 do not allow us to disentangle whether the evolution is in luminosity , in number density or in both . in sect . [ sect : lf_schechter ] we attempt to analyse this in a more quantitative way . the dashed grey curve in figure [ fig : lf ] marks the @xmath282 lf of @xcite measured with data from the galaxy and mass assembly ( gama ) survey . compared to our lowest redshift lf measurement , the @xmath282 lf is characterised by a steeper decay for @xmath283 . fully understanding the evolution of the lf from @xmath16 to @xmath282 goes beyond the scope of the present work . here we just speculate that a higher @xmath13 ratio at @xmath282 would increase the stellar mass of the brighter objects , generating consistency between the smfs at @xmath282 and @xmath16 . an increase of the @xmath35 ratio with cosmic time is indeed to be expected . the likely more recent star - formation activity at earlier cosmic times corresponds to higher luminosities at higher redshift ; assuming a constant stellar mass , this decrease of the luminosity with cosmic time corresponds to an increase of the @xmath35 with cosmic time . the stellar mass itself is also expected to grow with cosmic time , further increasing to @xmath35 towards @xmath282 . cdd @xmath177 & 2cstellar mass & 2c@xmath204 + bin & 2c@xmath284 & 2c(@xmath205 mpc@xmath206 dex@xmath207 ) + @xmath16 & 11.55 & @xmath285 + & 11.41 & @xmath286 + & 11.26 & @xmath287 + & 11.12 & @xmath288 + & 10.97 & @xmath289 + & 10.82 & @xmath290 + & 10.68 & @xmath291 + & 10.53 & @xmath292 + & 10.38 & @xmath293 + & 10.24 & @xmath294 + & 10.09 & @xmath295 + & 9.94 & @xmath296 + & 9.84 & @xmath297 + & 9.74 & @xmath298 + & 9.64 & @xmath299 + & 9.54 & @xmath300 + & 9.44 & @xmath301 + & 9.34 & @xmath302 + & 9.24 & @xmath303 + & 9.14 & @xmath304 + & 9.04 & @xmath305 + & 8.94 & @xmath306 + & 8.84 & @xmath307 + & . & . + @xmath91 & 10.68 & @xmath308 + & 10.38 & @xmath309 + & 10.09 & @xmath310 + & 9.84 & @xmath311 + & 9.64 & @xmath312 + & 9.44 & @xmath313 + & . & . + @xmath92 & 10.47 & @xmath314 + & 9.94 & @xmath315 + & 9.60 & @xmath316 + & . & . + @xmath17 & 9.84 & @xmath317 + we generated smf measurements taking advantage from the mass - to - light ratios we measured in section [ sect : ml_ratios ] . the feature that the @xmath13 does not decrease with luminosity turns out to be very useful in this . indeed , if the @xmath13 decreased instead of increasing , the same stellar mass could be obtained from a higher luminosity but lower @xmath13 or from a lower luminosity , but higher @xmath13 , and therefore the luminosity would not univocally determine the stellar mass . a non - decresing @xmath13 , instead , constitutes an injective mapping between luminosity and stellar mass : galaxies with higher luminosity will _ always _ have higher stellar mass , and galaxies with lower luminosity will _ always _ have lower stellar mass . taking into account the above considerations , we adopted the following very simple procedure . we assumed that the constant @xmath183 and the linear relation observed at @xmath16 ( eq . [ eq : ml_const ] and [ eq : ml ] ) were valid at all redshifts . the absolute magnitudes corresponding to the bin centers of the lfs were converted into stellar mass applying the relevant @xmath183 relation depending on the @xmath9 value ( see eq . [ eq : ml_const ] and eq . [ eq : ml ] ) . we then differentiated the two relations , solving for @xmath318 . the obtained values , specific for each @xmath5 bin , were used to rescale the lf normalisation , to take into account the change in units from mag@xmath207 to dex@xmath207 . our smf measurements are presented in figure [ fig : smf ] and table [ tab : smf ] . unsurprisingly , the @xmath16 smf covers a range in stellar mass wider than the @xmath84 smfs , for the same reasons we described for the lf . in figure [ fig : smf ] we also plot a compilation of smf measurements from the literature ( @xcite ) . at @xmath16 , starting from the low - mass end where the measurements are generally quite consistent with each other , the discrepancies increase with increasing stellar mass . one possible reason for the increased dispersion at higher masses is that galaxies constituting the low - mass end are mostly star - forming . their redshift can then be assessed through the location of the observed lyman break ( either from dropouts or photometric redshift selections ) . the massive end , instead , possibly also includes more evolved and/or dusty systems and it is therefore more sensitive to the degeneracy in identifying the observed break as either the balmer/4000 break or the lyman break . our @xmath16 smf determination is in good agreement with the smfs of @xcite , @xcite , @xcite , @xcite and @xcite . this is quite remarkable , since these smfs have been recovered from different selection techniques . specifically , @xcite and @xcite measurements are based on dropouts samples from _ hst_/wfc3 data ; the smf of @xcite was built from a @xmath60-detected photometric redshift sample over the candels / goods - s and candels / uds fields ; @xcite assembled a composite sample complementing a @xmath27-detected catalog from ultravista data with detections in irac @xmath7 m and @xmath51 m bands ; finally , @xcite measurements are based on a @xmath27-detected smf complemented by smf measurements from detections in irac @xmath51 m . the sample selection of both @xcite and @xcite relies on photometric - redshift measurements . on the other side , the normalisation of our smf is higher than @xcite and @xcite smfs ; both these measurements were obtained by converting the observed uv lf into smf through @xmath161 measurements . the discrepancy with our smf ( and the bulk of the other smf determinations ) could be due to a steeper @xmath161 relation found by @xcite and @xcite and consequent lower normalisation term . at the massive end ( @xmath319 ) , we observe a discrepancy between our @xmath16 smf and the corresponding measurements from the literature ( e.g. , @xcite ) . this discrepancy could , at least in part , be explained by our smf lacking any scatter in @xmath13 for a given @xmath320 . our stacking analysis , by construction , recovers median @xmath13 ratios , potentially excluding extreme cases such as very dusty / old systems with very high stellar masses . furthermore , because of the eddington bias , a distribution in the observed @xmath13 values for a specific luminosity would introduce a higher fraction of lower stellar mass objects scattered to higher stellar masses than the opposite , increasing the number density of the massive objects . at @xmath91 , the measurements of the low - mass end ( @xmath321 ) of the smf are characterised by a larger scatter than for @xmath322 . at higher stellar masses measurements are broadly consistent with each other mainly because of the large uncertainties on the number densities . our smf determination overlaps with the measurements of @xcite , @xcite , @xcite , @xcite and @xcite . however , it lies below the smf measurements of @xcite . at even higher redshift , the poisson uncertainties on the number densities start to be of the same order as those from cosmic variance . the small number of galaxies at @xmath92 and @xmath17 ( 10 and 2 galaxies , respectively ) generates large poisson uncertainties which result in a broad agreement among the different smf determinations . our smf overlaps with the measurements of @xcite , @xcite and @xcite . the large uncertainties make our smf measurements roughly consistent also with those of @xcite and @xcite . the shape of the lf and smf of galaxies is usually well described by a @xcite function over a wide range of redshifts ( @xmath323 , but see also e.g. @xcite ) . we therefore fitted our lf measurements with a schechter function : @xmath324 where @xmath325 is the characteristic magnitude , corresponding to the knee of the density distribution , @xmath326 is the faint end slope and @xmath327 is a global normalisation factor . we performed a best - fit to our lfs measurements using the levenberg - marquardt method . for the @xmath16 lf we left the three schechter parameters free to vary . since the faint - end slope of the @xmath76 lfs is poorly sampled , the best fits to the @xmath76 lfs were done fixing @xmath326 to the value obtained for the @xmath16 lf . the fit to the @xmath91 and @xmath92 lfs were done in three different configurations : both @xmath325 or @xmath327 as free parameters , with @xmath325 as unique free parameter and with @xmath327 as the only free parameter . we did not fit the @xmath17 lf leaving both @xmath325 or @xmath327 as free parameters . the resulting best - fit schechter functions for the case of pure luminosity evolution are marked in figure [ fig : lf ] with solid curves , while table [ tab : schechter ] lists the recovered parameters and their uncertainties . cccc redshift & @xmath327 & @xmath325 & @xmath326 + bin & ( @xmath205 mpc@xmath206 mag@xmath207 ) & ( mag ) & + + 4 & @xmath328 & @xmath329 & @xmath330 + & & & + 5 & @xmath331 & @xmath332 & @xmath333 + & @xmath334 & @xmath335 & @xmath333 + & @xmath336 & @xmath337 & @xmath333 + & & & + 6 & @xmath338 & @xmath339 & @xmath333 + & @xmath334 & @xmath340 & @xmath333 + & @xmath341 & @xmath337 & @xmath333 + & & & + 7 & @xmath334 & @xmath342 & @xmath333 + & @xmath343 & @xmath337 & @xmath333 + visual inspection of the best - fitting schechter function showed that , overall , there is a preference for a pure luminosity evolution against a pure density evolution . moreover , when both @xmath325 and @xmath327 were left free to vary , the values of @xmath327 were characterized by large uncertainties , making them consistent with the value of @xmath327 at @xmath16 , i.e. , corresponding to no evolution with redshift . on the contrary , the values of @xmath325 showed a more clear trend with redshift , increasing our confidence on the luminosity evolution . we note however , that these results potentially suffer from the limited coverage of @xmath325 . the luminosity evolution registered here is in contrast with the most recent measurements of the evolution of the uv lf at @xmath160 , where indication is found that the characteristic magnitude evolves very little with redshift , as most of the evolution seems to be driven by variation in the overall density ( see e.g. , @xcite ) . the lfs presented in our work are the first determination of the rest - frame optical lf at @xmath24 and therefore direct comparisons to previous estimates are not possible . however , recent works have recovered the measurement of the characteristic magnitude in the rest - frame @xmath3-band at @xmath16 ( @xcite ) or have studied the evolution of the lf up to @xmath344 in rest - frame optical bands close to @xmath3 ( @xcite ) . in the following paragraphs we will compare our determination of the @xmath16 characteristic magnitude to the estimates from the above three works . @xcite estimated the characteristic magnitude of the @xmath16 lf applying a correction based on the average @xmath345 $ ] color to the characteristic magnitude of the @xmath16 uv lf . the obtained value , @xmath346 ab , is @xmath347 mag fainter than what we found in this work . a possible reason for this large discrepancy is the lack of galaxies brighter than @xmath348 from the sample of @xcite , which instead have become accessible through the deep and wide area ultravista data . @xcite presented estimates of the rest - frame @xmath349-band lf at @xmath344 , obtained from a composite sample including wide - area data from the nmbs , fires , fireworks , hdfn , hudf and goods - s programs . the characteristic magnitude recovered from a maximum likelihood analysis is @xmath350 . this value is brighter than that found by @xcite , although it is still @xmath351 mag fainter than the estimate in our work ; nonetheless it is consistent at @xmath94 with our estimate , considering the associated uncertainties . our stacking analysis ( see figure [ fig : stack ] ) showed that the brighter galaxies have redder @xmath352 colors , reaching @xmath353 mag for the brightest stacks . using our stacked sed at @xmath154 , close to the value of the characteristic magnitude of our @xmath16 lf , we find a rest - frame color @xmath354 mag , which , applied to the @xmath355 , gives a @xmath356 mag , very close to our best - fit @xmath357 mag . @xcite measured the evolution of the rest - frame @xmath358 and @xmath359band lf up to @xmath360 using a composite sample of galaxies from the musyc , fires and goods - cdfs programs . from maximum likelihood analysis , the characteristic magnitude of the @xmath361 rest - frame @xmath358band lf was estimated to be @xmath362 mag . applying the same analysis adopted for the comparison to @xcite , we find a rest - frame color @xmath363 mag , and a corresponding @xmath364 mag , still consistent with our estimate at @xmath138 level . + we also performed a schechter fit to the smfs . however , the differential @xmath13 we measured at @xmath16 and applied to the @xmath76 lfs has the effect of _ stretching _ the original lfs . the consequence of this stretching , together with the limited number of measurements available at each redshift , is that the @xmath76 smfs do not show any robust evidence for the exponential cut at the massive end , preventing any reliable estimate of the characteristic stellar mass . therefore , we performed the schechter fit only on the @xmath16 smf , and obtained the following results : @xmath365 , @xmath366 mpc@xmath206 dex@xmath207 and @xmath367 . the corresponding best - fit schechter function is presented in figure [ fig : smf ] . the large uncertainties associated to the schechter parameters make them consistent with most of the measurements from the literature . the low number of massive galaxies from the exponential part of the schechter function , at the massive end , suffer from high uncertainties from both ( relative ) poisson noise and cosmic variance . the massive end of our smf is based on the ultravista sample , suggesting a more adequate coverage of massive galaxies . one of the main aim of this paper was to test the lf in rest - frame optical bands as a proxy for smf measurements at high redshifts ( @xmath24 ) . the compilation of smf measurements presented in figure [ fig : smf ] shows that for @xmath368 they are characterised by a scatter which can be as large as @xmath369 dex . this has immediate consequences on our understanding of more fundamental and global properties , like the evolution with cosmic time of the stellar mass density or the stellar - to - halo mass relation . the likely main reason for this large scatter is sample selection . in addition to this , further systematics may arise from our limited knowledge on some of the specific aspect characterising the stellar population of each galaxy , including contamination from emission lines ( see e.g. , @xcite ) . reliable stellar mass measurements require coverage of rest - frame optical / nir wavelengths , probed by _ spitzer_/irac at @xmath24 . in particular , current measurements of the low - mass end of the @xmath160 smfs rely on irac data over the goods - n / s fields as these are the only fields providing photometric coverage over a sufficient area and with the sufficient depth . the @xmath7 m and @xmath51 m bands , although characterised by deeper data than the @xmath31 m and @xmath8 m bands , are contaminated by nebular lines at @xmath370 . this means that robust stellar mass estimates need to rely on flux measurement in the two redder irac bands , i.e. , at @xmath31 m and @xmath8 m . through the @xmath13 we derived in the present work , the s / n cuts we applied to the flux in the irac @xmath31 m and @xmath8 m bands identify a range in stellar mass where their measurements can be considered reliable . the lf and smf from this work , then , constitute a _ skeleton _ of most of the current smf measurements at @xmath368 . this is visible in figure [ fig : smf ] : at @xmath91 and above the lowest stellar mass over which our smfs are defined is @xmath371 higher than most current smf determinations . specifically , this also means that stellar masses below our low - mass limits are necessarily based on either very low s / n measurements in the irac @xmath31 m and @xmath8 m bands or on ( still uncertain ) correction for nebular emission contamination , or a combination of the two . our analysis showed that smfs consistent with the average determinations from the literature could be recovered by applying a simple @xmath13 relation to the observed @xmath3-band lf , with stellar masses measured from common stellar population parameters ( e.g. , delayed exponential sfh , solar metallicity , chabrier imf ) . our simple transformation of the lf into smf was supported by the non - decreasing @xmath13 ratio for increasing luminosities observed at @xmath16 . a @xmath13 ratio decreasing with luminosity would correspond to lower mass galaxies in the bright end of the lf , which would _ remix _ the number densities in stellar mass bins . the ideal rest - frame band for this kind of studies is probably one for which the @xmath35 ratio would not depend on the luminosity , as any luminosity dependence could potentially hide effects from e.g. , sfh . our measurements of the @xmath13 relation suggest that , for a wide range in luminosity , they are consistent with a constant value . a log - linear relation arises at the bright- ( massive- ) end of the lf ( smf ) , suggesting that at these luminosities the @xmath3-band holds at some level the signature of the stellar population age and/or of the dust content . given the potential systematics on the smf measurements discussed above , lf estimates can provide a more robust basis for recovering the halo masses ( @xmath372 ) for high - redshift galaxies through abundance matching techniques ( e.g. , @xcite ) . to date , measurements of @xmath368 lfs are mostly available in the rest - frame uv . the adoption of uv lfs in the @xmath373 estimates provides information on the relative importance of star formation processes ( e.g. , gas cooling , stellar ejections , sfr timescales ) versus the hierarchical growth of the dark matter halos ( @xcite ) . however , stellar masses are likely to be more strongly correlated with the halo masses than uv luminosities . the schechter fits performed in sect . [ sect : lf_schechter ] suggest that the evolution of the @xmath3-band lfs can be accounted for by an increase of luminosity with cosmic time uniformly across luminosities at a given redshift . since we showed that the @xmath3-band lf is a reasonable proxy for the smf , it is tempting to analyse the evolution of the rest - frame optical lf obtained in the present work in terms of the evolution of the dark - matter halo mass function ( hmf ) . we performed a first analysis as follows . we applied a simple abundance matching technique ( @xcite ) consisting in matching the cumulative number density of the lf to that of @xcite hmf obtained from ` hmfcalc ` @xcite and recovered the evolution in mass of the halo mass function at @xmath280 relative to @xmath16 . given the rapid evolution of the hmf in this range of redshift , we adopted the hmfs at @xmath374 and @xmath375 , corresponding to the ( median ) photometric redshifts of the stacked seds . the hmfs assumed @xmath376 . we found a relative displacement in halo mass of 0.76 , 0.52 and 0.35 dex , corresponding to @xmath377 , @xmath378 and @xmath379 mag from @xmath380 and @xmath381 to @xmath16 , respectively . figure [ fig : lf4hmf ] shows the result of applying the above offsets to the @xmath382 of the @xmath16 schechter parameterization . the solid blue curve marks the best - fit schechter function at @xmath16 , while the dashed curves represent the @xmath16 lf rigidly shifted by the corresponding amount at @xmath76 . the agreement between the predicted and observed lf is very good at all redshifts , suggesting that the @xmath3-band lf could trace the evolution of the hmf . in this section we discuss the evolution of those halos associated to a constant cumulative number density of @xmath384 mpc@xmath206 over @xmath385 . this value corresponds to the cumulative number density of @xmath382 galaxies at @xmath16 ( @xmath386 mag ) and it allows us to recover the corresponding absolute magnitude up to @xmath17 with strongly reduced dependence ( i.e. , @xmath387 mag ) on the extrapolation of the lfs to magnitudes fainter than actually observed . table [ tab : hmf_fix_cnd ] lists a compilation of the values of the main parameters recovered with our analysis . [ cols="^,^,^,^,^,^,^,^,^ " , ] the abundance matching performed through cumulative number density implicitly assumes that each halo contains one and only one galaxy and that halos of the same mass contain galaxies of the same @xmath3 luminosity ( @xmath4stellar mass ) . indeed , recent measurements have shown that the scatter between halo mass - stellar mass relation is quite small , @xmath388 dex ( e.g. , @xcite , but see also @xcite who found scatter of up to 0.32 dex ) . the cumulative number densities of the lfs were computed adopting the schechter parameterization presented in sect . [ sect : lf_schechter ] . for the lfs at @xmath76 , we adopted the best - fit schechter functions obtained when the characteristic magnitude was assumed to be the only free parameter of the fit , coinciding with the case of pure luminosity evolution . the values of the absolute magnitudes we obtain from our procedure are @xmath389 mag for the @xmath16 , @xmath91 , @xmath92 and @xmath17 cases , respectively . these values are consistent within @xmath138 with the characteristic magnitudes of our schechter fit . this is not surprising , considering the pure luminosity evolution of the schechter fits themselves . the matches to the cumulative number density performed on the hmfs resulted in halo masses @xmath390 for the @xmath16 , @xmath91 , @xmath92 and @xmath17 cases , respectively . these displacements correspond to an evolution in the halo mass of @xmath382 galaxies of @xmath391 , and @xmath392 dex from @xmath393 to @xmath16 , respectively . we note , however , that a rigid displacement in mass of the hmf is sufficient to reproduce the hmf evolution at these redshifts only for halo masses @xmath394 ; at lower halo masses , the displacement in mass must be coupled to a steepening with redshift of the low - mass end slope . we can now use the above results on the evolution of the luminosity and of the halo mass to recover the evolution with redshift of the light - to - halo mass for galaxies at fixed cumulative number density . combining the two we obtain @xmath395 in units of @xmath396 . these values are also presented in the left panel of figure [ fig : hmf_fix_cnd ] and suggest a mild increase with redshift ( a factor @xmath397 ) , although the large uncertainties make them consistent with a constant value across the 800 myr of cosmic time from @xmath16 to @xmath17 . using the results on the @xmath13 from sect . [ sect : ml_ratios ] , we can convert the @xmath398 into @xmath399 . the result of this is shown in the right panel of figure [ fig : hmf_fix_cnd ] . the corresponding values are listed in table [ tab : hmf_fix_cnd ] . the stellar - to - halo mass ratio does not present any significant evolution with redshift . in the same panel we convert the @xmath399 into the integrated star - formation efficiency ( isfe i.e. , @xmath399 in units of @xmath400 - @xcite , and equivalent to the stellar baryon fraction - @xcite ) , using @xmath401 @xcite . our measurements are consistent with the isfe being constant with redshift . we stress here that this result refers to @xmath402 and it does not exclude the existence of evolution with redshift at lower halo masses . we defer a more complete analysis on the dependence of the @xmath399 ratio with halo mass to a future work . furthermore , our samples at @xmath403 are entirely based on lbg selection . if non - negligible numbers of redder ( dustier / more evolved ) galaxies exist at these epochs , they would affect smf ( and likely its massive end , e.g. @xcite ) and , consequently , the recovered @xmath372 . matching galaxies at a constant cumulative number density , however , does not consider the effect of major mergers in the galaxy ranking . we therefore repeated the same analysis using a cumulative number density evolving with redshift following the recipe of @xcite . the results are listed in table [ tab : hmf_fix_cnd ] , and plotted as red circles in figure [ fig : hmf_fix_cnd ] . no significant difference with the constant cumulative number density match is observed . we also note that the values for the @xmath17 bin rely on the extrapolation of the lf to luminosities below those currently probed by our sample . those measurements should then be treated with caution . in the right panel of figure [ fig : hmf_fix_cnd ] we also plot recent estimates of the @xmath399 from the literature : @xcite , ( * ? ? ? * converted to a @xcite imf by applying a factor 0.55 ) and @xcite . @xcite applied the measurements of the two - point correlation function to a halo occupation model to recover the halo mass of samples of galaxies at @xmath404 with spectroscopic redshift from the vimos ultra deep survey ( vuds - @xcite ) . @xcite measured the evolution of the @xmath399 from abundance matching the @xmath405 uv lf . @xcite recovered @xmath399 from the clustering of lbgs selected at @xmath405 from a variety of programs , including candels , the _ hubble _ frontier fields ( pi : j. lotz ) and subaru hyper - suprime cam subaru strategic program ( pi : s. miyazaki ) . at face value , our measurements are consistent with those of @xcite at @xmath16 and with those of @xcite at @xmath406 ; however , they are inconsistent with those of @xcite over the full range of redshift , and with those of @xcite at @xmath407 . the @xmath399 measurements presented in the right panel of figure [ fig : hmf_fix_cnd ] were obtained from a variety of methods , and ultimately refer to different @xmath372 estimates , making the straight comparison difficult to interpret . specifically , our measurements of the halo mass are based on a constant cumulative number density match with @xmath408 at @xmath409 , respectively . the increase with redshift of the @xmath399 of @xcite refers to a _ fixed _ @xmath410 across @xmath405 . @xcite report an increase with redshift of the isfe @xmath411 . however , this trend is most likely driven by the point at @xmath16 , and originated by limited evolution of the uv lf between @xmath16 and @xmath91 ( @xmath412 mag ) observed by @xcite , in contrast to the large luminosity evolution ( @xmath413 mag ) observed in our work over the same redshift interval . this , together with the constant characteristic magnitude of the uv lf ( i.e. , @xmath414 constant sfr ) assumed as criterion for the abundance matching , generates a reference cumulative number density decreasing with redshift , and halo masses @xmath415 . we note here that our measurements are consistent with those of @xcite at @xmath406 , i.e. , where the @xmath372 recovered by the two teams are more similar , while they are inconsistent at @xmath407 , where the @xmath372 differ . finally @xcite estimates refer to a diversity of halo masses and redshift ranges : @xmath416 for @xmath417 , respectively . our finding of @xmath399 independent of redshift is also qualitatively in agreement with recent estimates of the galaxy bias , observed to be nearly constant over @xmath418 ( @xcite ) , although a measurement at @xmath17 from the same work seems to suggest a potential change of the star formation efficiency at earlier epochs . the picture is not settled even from a theoretical perspective . indeed , some of the models predict a _ decrease _ with redshift of the @xmath399 for a fixed @xmath372 ( e.g. , @xcite ) . other models , instead , result in @xmath399 to be insensitive to redshift ( e.g. , @xcite ) . finally , semi - empirical models have been able to reproduce the evolution of the uv lf from @xmath20 to @xmath419 under the assumption that , for star - forming galaxies , the @xmath399 depended on @xmath372 but not on redshift ( @xcite ) . the main aim of this work was to measure the rest - frame @xmath3-band luminosity function ( lf ) of field galaxies and to study its evolution at @xmath368 . the rest - frame @xmath3 band was selected for three reasons : 1 ) it is not contaminated by strong emission from nebular lines ; 2 ) light in this wavelength range is dominated by lower mass , long - living stars ; and 3 ) it can be probed up to @xmath6 using the current _ spitzer_/irac data . these characteristics suggest its use as a proxy for stellar mass at high redshift , minimizing the potential systematic effects that can affect stellar mass measurements . we therefore assembled samples of lyman break galaxies ( lbgs ) at @xmath16 , @xmath91 , @xmath92 and @xmath17 , selected over the goods - n and goods - s fields . the @xmath16 sample was complemented by galaxies with photometric redshifts @xmath78 extracted from a 37-band far - uv - to-@xmath8 m @xmath27-detected photometric catalog based on ultravista dr2 . the larger @xmath16 co - moving volume provided by the ultravista data allowed us to gain statistics on the rarer more luminous and/or redder galaxies . the goods - n / s sample takes advantage from the recently released full depth irac maps @xcite , obtained from the combination of all the irac programs carried out so far over these fields , namely igoods , iudf , goods , ers , s - candels , seds and ufd2 . these maps reach a depth of @xmath420 mag and @xmath32 mag in the @xmath51 m and @xmath31 m bands respectively ( @xmath56 diameter aperture , @xmath33 ) , although the coverage is highly inhomogeneous . similarly , the ultravista catalog benefits from irac @xmath7 m and @xmath51 m mosaics which combine the s - cosmos , s - candels and splash programs and reach a depth of @xmath421 mag ( @xmath56 diameter aperture , @xmath33 ) . we further selected our sample based on the s / n in the irac band or bands closer to the rest - frame @xmath3 band . specifically , the final @xmath16 sample was selected to have @xmath83 in the @xmath51 m while the @xmath422 samples were selected to have @xmath423 in the inverse - variance weighted combination of s / n in the @xmath31 m and @xmath8 m bands . our final composite sample included 2098 , 72 , 10 and 2 objects , for the @xmath79 and 7 redshift bins , respectively . although the @xmath3 band is covered by the @xmath31 m and @xmath8 m irac data up to @xmath6 , we do not register any lbg galaxy at @xmath6 which also satisfies our selection criteria on the s / n of irac fluxes . our main results are as follows : 1 . at @xmath16 and for absolute magnitudes @xmath9 fainter than @xmath424 ab , galaxies follow a linear relation on the @xmath140 plane , with slope @xmath425 . this correlation breaks at @xmath426 ab : the @xmath136 of these galaxies covers the full range of values observed for galaxies with fainter @xmath9 ( figure [ fig : uvopt ] ) . we performed stacking analysis and measured the @xmath13 of galaxies segregated according to their redshift and absolute magnitude @xmath9 . the @xmath13 at @xmath16 is independent of @xmath9 for @xmath427 ; at brighter @xmath9 the @xmath13 increases with luminosity following a power - law . the @xmath13 at @xmath76 are consistent with those observed at @xmath16 , although the associated large uncertainties may hide a different behaviour ( figure [ fig : stack ] and [ fig : ml ] ) . 3 . we computed the lf in the rest - frame @xmath3 band , using the @xmath267 estimator , in four different redshift bins : @xmath16 , @xmath91 , @xmath92 and @xmath17 . the lf shows evolution from @xmath17 to @xmath16 . schechter fits to the @xmath267 lf marginally prefer pure evolution in luminosity over a pure evolution in density ( figure [ fig : lf ] ) . 4 . the non - decreasing @xmath13 with luminosity ( corresponding to an injective mapping ) allowed us to apply a simple conversion from luminosity to stellar mass . we therefore converted our lf measurements into smf using the @xmath13 recovered from the stacking analysis at @xmath16 . the obtained smfs are consistent with the average smf determination from the literature . despite the relaxed s / n cuts in irac flux applied to our samples , the lower stellar mass over which we recover our smfs is @xmath428 larger than typical lower limits from the literature ( figure [ fig : smf ] ) . evolution in the halo mass relative to @xmath16 recovered from abundance matching the halo mass functions reproduces the luminosity evolution of the lf at @xmath24 ( figure [ fig : lf4hmf ] ) . the stellar - to - halo mass ratio at fixed cumulative number density shows no strong evidence for evolution with redshift over @xmath429 ( figure [ fig : hmf_fix_cnd ] ) . the above results allow us to draw the following conclusions : 1 . the current depth of _ spitzer_/irac data revealed to be sufficient to probe the regimes in rest - frame uv luminosities both where the rest - uv luminosities are correlated with stellar mass and where they are not . the existence at @xmath16 and @xmath91 of lbgs luminous in the rest - frame @xmath3 band and spanning a broad range in uv luminosities suggests that the adoption of the uv lf for smf estimates may be affected by systematics . specifically , samples of galaxies selected to have a narrow range in uv luminosities potentially include a combination of high and low mass objects , and ultimately can introduce an over - estimate of the low - mass end slope and an underestimate of the densities at the high - mass end . the higher values of the lower stellar mass bin in our smfs compared to recent determination from the literature , arising from the s / n cuts applied to the irac fluxes , suggests that current low - mass end of the smfs at @xmath24 might be based on low s / n flux measurements ( @xmath430 upper limits ) in the observed irac bands most sensitive to the stellar mass ( i.e. , @xmath31 m and @xmath8 m ) . higher s / n measurements are available from @xmath7 m and @xmath51 m data . however these bands at @xmath52 are contaminated by nebular emission which can potentially bias the stellar mass estimates , given our still limited knowledge on the emission line intensities of high - redshift star forming galaxies . the rest - frame @xmath3 band lf can be a valid proxy for smfs and hmf measurements at @xmath24 . the nearly flat dependence of the @xmath13 on @xmath9 increases this confidence . this work is largely based on data from the cryogenic programs of _ spitzer_/irac . while the depth of the @xmath7m- and @xmath51m- band data can still be improved through non - cryogenic programs , the sensitivity of jwst / miri provides the only opportunity for increasing the depth at wavelengths @xmath431 m , necessary for improving current estimates of stellar masses at @xmath52 . + ms would like to thank adriano fontana for constructive discussions . ms and rb are grateful to karina caputi , yuichi harikane , michele trenti and stuart wyithe for helpful feedback on an advanced draft of this manuscript . this work is based on data products from observations made with eso telescopes at the la silla paranal observatory under eso programme i d 179.a-2005 and on data products produced by terapix and the cambridge astronomy survey unit on behalf of the ultravista consortium . this work is based on observations taken by the candels multi - cycle treasury program with the nasa / esa hst , which is operated by the association of universities for research in astronomy , inc . , under nasa contract nas5 - 26555 . this work is based on observations taken by the 3d - hst treasury program ( go 12177 and 12328 ) with the nasa / esa hst , which is operated by the association of universities for research in astronomy , inc . , under nasa contract nas5 - 26555 . after applying the s / n cuts described in sect . [ sect : sample_assembly ] , we further cleaned our sample , excluding those objects satisfying any of the following conditions : 1 ) the contribution to the @xmath31 m and @xmath8 m flux from neighbouring objects is excessively high ; 2 ) the source morphology is very uncertain or confused making irac photometry undetermined ; 3 ) the source is detected at x - rays wavelengths , suggesting it is a lower redshift agn ; 4 ) the source is at higher redshift , but its sed is dominated by agn light ; 5 ) lbgs with a likely @xmath93 solution from photometric redshift analysis . in the following paragraphs we will describe in more detail the above criteria and their effects on the sample size . the broad irac psf together with the unprecedented photometric depth of the irac mosaics in the goods - n , goods - s and ultravista fields can result in flux measurements potentially affected by contamination from brighter nearby objects . the procedure we adopted for the flux measurements already deals with this problem by cleaning each source from its neighbours before performing the photometry . however , in some cases the flux at the position of the object of interest mostly comes from the bright neighbours , resulting in potentially very uncertain flux measurements . we therefore opted to further clean our sample by applying a cut on the maximum fraction of flux from neighbours contributing to the flux of each object before the neighbour - cleaning process . specifically we excluded from our sample those sources whose neighbours were contributing more than 65% to the total flux at the position of each source in our sample . we also visually inspected the cutouts from the irac photometry and further excluded those sources showing a residual contamination from bright nearby sources . in this step we removed 280 galaxies ( 211/69 , for goods - n / s and ultravista , respectively ; of the 211 goods galaxies , 3 were at @xmath92 and 1 at @xmath17 ) . in section [ sect : completeness ] we describe the monte carlo simulation we implemented to evaluate the selection effects introduced by the above selection criteria in a statistical way . we visually inspected the cutouts of the goods - n / s sample in the wfc3/@xmath60 band and of the ultravista sample in the acs / f814w , and excluded those objects with doubtful morphology , as e.g. , it was the result of two or more distinct objects or the deblending from sextractor was deemed inconsistent . furthermore , the visual inspection also allowed us to identify and exclude objects with point - source morphology as either potentially agn dominated or brown dwarfs contaminants . this was particularly important for the ultravista sample , since the @xmath27-band detection image is characterized by a psf fwhm @xmath432 , much broader compared to that of acs or wfc3 ( fwhm @xmath433 and @xmath434 , respectively ) . this class of objects are subject to very inaccurate photometry , redshift classification and/or luminosity measurement . through the above criteria we excluded 53 objects from the goods - n / s sample ( 1 at @xmath17 ) and 63 objects from the ultravista sample . successively , we crossmatched our sample to catalogs of x - ray sources in the goods - n / s @xcite and cosmos fields @xcite , and excluded all the matching sources as these are potential lower - redshift agn contaminants . we identified 34 sources with an x - ray counterpart matching our initial sample , most of which at @xmath16 ( 29 ) , and 4 at @xmath91 . furthermore we visually inspected all the observed seds to exclude either objects with very red , power - law like rest - frame optical / nir slopes which could be signature of type-1 agn . in this step we flagged and removed from the sample a total of 74 sources ( 20/54 ) . finally , we run eazy @xcite on the sample of lbgs , and excluded those galaxies with @xmath435 , sources with prominent secondary lower-@xmath177 solution or inconsistent sed . through this step we excluded 109 sources ( 4 of which at @xmath92 ) . the final sample consists of 2098 galaxies at @xmath16 ( 1680 from the lbg sample and 418 from the ultravista sample ) , 72 at @xmath91 , 10 at @xmath92 and 2 objects at @xmath17 . in figure [ fig : seds_z5 ] we present the seds of the 12 most luminous galaxies included in the @xmath91 sample , while figures [ fig : seds_z6 ] and [ fig : seds_z7 ] show the full sample of galaxies at @xmath92 and @xmath17 , respectively . + most of the galaxies are characterised by small , compact sizes in the @xmath60 . noticeably , gsdi-2244050099 , at @xmath92 , is among the most luminous galaxies of the full sample , including @xmath16 . its apparent size is larger than the average size of the galaxies in the @xmath436 and @xmath1 samples . careful inspection of the wfc3/@xmath60 cutout does not show any indication of clumpiness . however , in acs / f184w we observe 2 possible components , separated by @xmath437 ( @xmath438 kpc at @xmath439 ) . we opted for including it in our sample given the small separation between the two components visible only in the acs data , the consistency of the sed and the fact that the irac flux appears to be centered at the position of the brighter component . this object constitutes the unique element of the highest luminosity bin of the @xmath92 lf . noteworthy , even assuming a factor 2 overestimate of the irac flux , the resulting absolute magnitude would still be consistent with the highest luminosity bin .	we present the first determination of the galaxy luminosity function ( lf ) at @xmath0 and @xmath1 in the rest - frame optical at @xmath2 nm ( @xmath3 band ) . the rest - frame optical light traces the content in low - mass evolved stars ( @xmath4stellar mass - @xmath5 ) , minimizing potential biases affecting measurements of @xmath5 : it is less affected by nebular line emission contamination , it is less sensitive to dust attenuation models , its measurement does not rely on stellar population models , and it can be probed up to @xmath6 through _ spitzer_/irac . our analysis leverages the unique full depth _ spitzer_/irac @xmath7m - to-@xmath8 m data over the candels / goods - n , candels / goods - s and cosmos / ultravista fields . we find that at absolute magnitudes @xmath9 fainter than @xmath10 mag , @xmath9 linearly correlates with @xmath11 . at brighter @xmath9 , @xmath11 presents a turnover , suggesting that the stellar mass - to - light ratio @xmath12 could be characterised by a very broad range of values at high stellar masses . median - stacking analysis recovers a @xmath13 roughly independent on @xmath9 for @xmath14 mag , but exponentially increasing at brighter magnitudes . we find that the evolution of the lf marginally prefers a pure evolution in luminosity over a pure evolution in density , with the characteristic luminosity decreasing by a factor @xmath15 between @xmath16 and @xmath17 . direct application of the recovered @xmath13 generates stellar mass functions consistent with average measurements from the literature . measurements of the stellar - to - halo mass ratio at fixed cumulative number density show that it is roughly constant with redshift for @xmath18 . this is also supported by the fact that the evolution of the lf at @xmath19 can be accounted for by a rigid displacement in luminosity corresponding to the evolution of the halo mass from abundance matching .
according to all attempts to build a solution to the solar neutrino anomaly based on the resonant spin - flavor precession mechanism @xcite , when one assumes a neutrino magnetic moment of order @xmath1 , which is slightly below its present experimental limit @xcite , an average magnetic field of order @xmath2 g @xcite is required to conciliate the present solar neutrino experimental data @xcite-@xcite and solar neutrino theoretical predictions @xcite-@xcite . using these typical values for the solar magnetic field we observed @xcite that magnetohydrodynamics @xcite naturally predicts the existence of solar magnetic oscillations which period is of the order of 1 to 10 days . resonant spin - flavor precession mechanism generates a solution to the solar neutrino anomaly only if neutrinos are very sensitive to the magnetic field as well as matter density profiles inside the sun and , in particular , are very sensitive also to the magnetohydrodynamics fluctuations of these quantities . therefore we expect that these fluctuations will generate an accordingly fluctuating experimental neutrino counting rate . the corresponding observable signal of this phenomenon is a periodical time fluctuation of the solar neutrino flux , in particular , in the high energy portion of the solar neutrino spectrum , detected in present real time experiments @xcite . these effects can be thought as a test to the resonant spin - flavor precession solution to the solar neutrino problem since they can clearly be distinguished from other sources of time modulation in the solar neutrino observations like the effect of seasonal variation of the sun - earth distance on vacuum oscillations , msw effects at the earth , and solar magnetic activity related to the appearance of sunspots on the solar surface . in general , magnetohydrodynamics @xcite predicts two different kinds of stable fluctuations . alfvn waves , when transverse oscillations of the fluid with respect to the magnetic field and the propagation vector are present and , therefore , no fluctuation of matter density or pressure is observed . and magnetosonic waves , which present longitudinal oscillations of density with respect to the propagation vector . in reference @xcite we analyzed the effects of purely alfvn waves where only magnetic fluctuations affects the solar neutrino survival probability . nevertheless , if matter density perturbations are present in the way described by magnetohydrodynamics , their effects can be important for the relevant high energy neutrinos observed in real time experiments @xcite . there are two main reasons which can be evoked to justify this . the presence of matter suppresses , in general , the neutrino spin - flavor precession mechanism , once that matter appears in the diagonal entries of the evolution matrix , which are spin - flavor conserving . that is why the presence of a resonance , a position along the neutrino trajectory where the matter effects are cancelled by a term proportional to the squared mass difference divided by the neutrino energy , is important for the neutrino spin - flavor conversion . therefore , fluctuations of the matter density can generate modifications in this resonance cancellation leading to important variations on the final probability of neutrino spin - flavor precession . the second reason to consider the effect of density fluctuations is the peculiar behavior of the adiabaticity parameter in the resonant spin - flavor precession phenomenon . in fact , several authors discussed the effect of density variations on the solar neutrino matter enhanced transitions , the msw effect @xcite . in this context , they conclude that the effects of density perturbations on transitions of solar neutrinos may be substantial if these perturbations occur in the central region of the sun and therefore only low energy neutrinos are sensitive to such effects since their resonance layer occurs deeper in the sun . note however that this conclusion can not be straightforward invoked here . while the adiabaticity increases with increasing resonance density for the msw effect , it is inversely proportional to the resonance density for the spin - flavor precession phenomenon and the effects of density perturbations in the solar convective zone may be important . this low density region is relevant for high energy solar neutrinos since a resonance for such neutrinos can be found there . consequently high neutrinos observed in real time experiments can be sensitive to matter fluctuations in this region . in this paper we investigate the impact of matter density fluctuations induced by sonic magnetohydrodynamics waves on observable time fluctuations of high energy solar neutrino flux detected in real time experiments and compare these effects with consequences from purely magnetic ( alfvn ) fluctuations . we conclude that both magnetic and density fluctuations can be equally relevant , although matter density fluctuations are much more constrained from helioseismology observations . taking into consideration such constraints , we still expect to find periodical time fluctuations in solar neutrino counting rate due to magnetohydrodynamics matter density waves of order of 10% of the total number of events . we are assuming a non - vanishing neutrino transition magnetic moment . in this case , the interaction of neutrinos with a magnetic field will generate spin - flavor precession which is given by the evolution equations @xcite @xmath3 where @xmath4 ( @xmath5 ) is the left ( right ) handed component of the neutrino field , @xmath6 is their squared mass difference , @xmath7 is the neutrino energy , @xmath8 is the fermi constant , @xmath9 is the electron number density distribution and @xmath10 is the transverse component of the magnetic field . finally we have @xmath11 for majorana neutrinos , in which case the final right - handed states @xmath5 are active nonelectron antineutrinos . for dirac neutrinos , @xmath12 , in which case the right - handed final states are sterile nonelectron neutrinos @xcite . in this paper we will assume majorana neutrinos . note however that the multiplicative factor 11/10 which has to be included in @xmath13 for dirac neutrinos does not lead to important alterations in our conclusions , which are , in this way , valid for majorana or dirac neutrinos . from eq . ( [ motion ] ) we observe that the neutrino spin - flavor precession mechanism depends on both the magnetic field @xmath10 as well as the matter density distribution @xmath9 along the solar neutrino trajectory . these quantities are affected by magnetohydrodynamics fluctuations . in order to describe them , we consider the linearized magnetohydrodynamics equations @xcite , following the same steps of reference @xcite , where magnetohydrodynamics spectrum is generated by small displacements @xmath14 from an equilibrium configuration . we are particularly interested in the magnetic and matter density fluctuations generated by the displacement @xmath14 given , respectively , by @xcite @xmath15 where @xmath16 is the equilibrium magnetic field configuration assumed to be the one relevant to the solar neutrino problem described in reference @xcite and @xmath17 is the solar matter distribution calculated in the standard solar model @xcite-@xcite . a crucial region of the magnetohydrodynamical spectrum is the continuum region which is associated with singularities of the hain - lst equation @xcite . this happens when the frequency of the magnetohydrodynamical fluctuation is equal to @xmath18 where @xmath19 is the pressure and @xmath20 is the ratio of specific heats . @xmath21 and @xmath22 define the alfvn and slow continuum regions in the magnetohydrodynamical spectrum @xcite . magnetic waves and matter density fluctuations associated with these frequencies are called localized modes since they present the interesting feature of been highly peaked around the position , @xmath23 , where the singularity occurs @xcite . in order to numerically overcome the singularities associated with the continuum spectra , a resistivity layer in the position of the singularity is introduced @xcite . the width of this layer is directly related with the width @xmath24 of the localized magnetic or matter density fluctuation which can be estimated @xcite : @xmath25 where @xmath26 is the vacuum permeability . in our particular scenario , @xmath24 can achieve 10@xmath27 ( normalized by the solar radius ) . to analyze the consequences of the localized waves on the solar neutrino observations , we have now to define the amplitude of possible magnetic and matter density fluctuations . the solar magnetic field is relatively free to fluctuate since the magnetic pressure @xmath28 is negligibly small when compared with the dominant gas pressure @xmath19 @xcite if we consider the matter density distribution @xmath17 predicted by the standard solar model and the magnetic field strength of order of those ones required to solve the solar neutrino anomaly @xcite . in fact , in this case , @xmath29 varies from approximately @xmath30 in the central regions of the sun to order @xmath31 close to the solar surface . from this argument , the magnetic field can be as large as 10@xmath32 g in the solar core or 10@xmath33 g in the solar convective zone . more stringent bounds on the magnetic field in the convective zone are found in refs . @xcite where the discussion is based on the non - linear effects which eventually prevents the growth of magnetic fields created by the dynamo process . by equating the magnetic tension to the energy excess of a sinking element at the bottom of the convective zone , schmitt and rosner @xcite obtained few times @xmath34 g as an upper bound . therefore fluctuations of the solar magnetic field of the same order of magnitude of the magnetic field used in reference @xcite can be found inside the sun . despite this fact , they can not be arbitrarily large when we are considering the linearized magnetohydrodynamics equations @xcite . this implies that the displacements @xmath14 must be small , @xmath35 , such that the non - linear terms can be neglected . this implies that @xmath36 . the error associated with this approximation is @xmath37 . the maximum possible value for the ratio @xmath38 is related with a clear statistical distinction between the maximum and the minimum value of the perturbed magnetic field , which is given approximately by @xmath39 ( in units of @xmath40 ) . to have a minimum 2-@xmath40 distinction between the maximum and minimum magnetic field , we must have a maximum value of the perturbation @xmath41 . something very different happens with possible matter density fluctuations . in fact , they are very constrained by helioseismology observations . the largest density fluctuations @xmath42 inside the sun are induced by temperature fluctuations @xmath43 due to convection of matter between layers with different local temperatures . an estimate of such effect is presented in reference @xcite and gives @xmath44 where @xmath45 is the nucleon mass , @xmath46 is the gravity acceleration and @xmath47 ( @xmath48 is the solar radius ) is a numerical factor coming from the approximately exponentially decreasing standard matter density distribution inside the sun @xcite . since @xmath49 is not in conflict with helioseismology observations @xcite , taking @xmath50 , we have to consider density fluctuations @xmath51 smaller than 10% . in fact , in an accurate analysis of helioseismology consequences on matter density fluctuations @xcite it was concluded @xmath51 can be very large ( larger than 10% ) only for very inner parts of the sun ( @xmath52 ) as well as for very superficial regions ( @xmath53 ) . for @xmath54 , @xmath51 decreases approximately linearly and achieves its smaller value 2% in @xmath55 . finally in the region where @xmath56 , @xmath51 is approximately 5% . we impose these constraints as boundary conditions for the amplitudes of density fluctuations we will consider in the following . finally we can analyze the consequences of the localized magnetohydrodynamics modes on the solar neutrino flux solving the neutrino evolution equations when a non - vanishing neutrino transition magnetic moment @xmath57 is assumed . we will adopt here a phenomenological approach , in close analogy to what was done in reference @xcite . we will assume that magnetohydrodynamics introduces gaussian shaped fluctuations which will be added to the equilibrium configuration of both matter density and/or magnetic field profile . this gaussian perturbation is centered in @xmath23 , with width @xmath24 . for the matter density fluctuation we have : @xmath58 \sin\left [ w(r_s)t \right ] , \label{rhopert}\ ] ] where @xmath59 is the fluctuation amplitude normalized to the value of the standard matter density @xmath60 calculated in the position of the singularity @xmath23 and @xmath24 is the width of the fluctuation . the frequencies @xmath61 or @xmath62 are given in equation ( [ continua ] ) and introduce a periodical time modulation on the standard matter density profile . similarly , our assumption for the magnetic fluctuations is obtained from the above equation substituting @xmath63 , i.e. , the standard matter density by the equilibrium field profile . therefore the final matter density ( as well as the magnetic field profile ) will be given by some equilibrium profile summed to the perturbation shown in eq . ( [ rhopert ] ) . we will assume the parameter @xmath59 and @xmath64 varying from 0 to 0.05 , which means that the matter density as well as the magnetic field fluctuate around their equilibrium values with maximum amplitude around 5% of this value . these fluctuations will generate an according time fluctuation of the neutrino counting governed by the evolution equations ( [ motion ] ) . in figure 1 we show the periods of the magnetic and matter density fluctuations associated with the continuum spectra given in ( [ continua ] ) for several magnetic field profiles given by the eqs . ( [ akhmedov])-([prof2 ] ) . we observe that for the considered magnetic fields typical periods vary from @xmath79 to @xmath80 days . time fluctuations of solar neutrino observations presenting periods of this order of magnitude constitute the most important signal of the existence of solar magnetohydrodynamics fluctuations and a crucial test for the resonant spin - flavor neutrino precession mechanism to the solar neutrino problem @xcite . the results of our analysis are shown in figure 2 . the fluctuations of the magnetic and matter density induce fluctuations in the survival probability @xmath81 , which is calculated solving the eqs . ( [ motion ] ) . in this figure we present the amplitude @xmath82 of the fluctuations of the survival probability as a function of @xmath83 , for some values of @xmath84 , the position of the localized mode . we show the phenomenon for @xmath85 , 0.9 and for the value of @xmath84 which gives the maximum probability amplitude , usually around @xmath86 . we adopted the magnetic field profile given in eqs . ( [ akhmedov ] ) and ( [ prof1 ] ) , with @xmath87 and @xmath88 . other magnetic field profiles shown in eqs . ( [ akhmedov ] ) , ( [ prof1 ] ) and ( [ prof2 ] ) generate very similar consequences of magnetohydrodynamics fluctuations on the solar neutrino survival probability and we will not show them here . we assumed @xmath59 and @xmath64 in eq . ( [ rhopert ] ) , i.e. , the relative amplitude of matter density and magnetic field profile fluctuations , respectively , to vary from 0 to 5% as it is indicated in figure 2 . also , vertical lines indicate the value of log@xmath89 corresponding to a resonance coinciding with the indicated singularity @xmath84 . we can observe that the fluctuations of the survival probability are maxima near the regions of the resonances and are well localized in @xmath83 for @xmath90 . we verified also that , for fixed values of @xmath84 and @xmath83 , the corresponding probability amplitude varies linearly with the magnitude of the amplitude of the matter density @xmath91 and/or magnetic field fluctuation @xmath92 . for instance , when @xmath93 , the maximal probability fluctuation corresponding to the magnetic fields given in eq . ( [ prof1 ] ) occurs in @xmath94 and can be written as @xmath95 for @xmath96 , @xmath97 for @xmath87 and @xmath98 for @xmath99 . considering the exponential behavior for the magnetic field in the convective zone , the maximal amplitude of survival probability occurs in @xmath100 and presents the following linear behavior with @xmath59 : @xmath101 . therefore , given the results of figure 2 , one can easily infer the amplitude of the survival probability for other values of @xmath91 and @xmath92 . note that this approximately linear behavior is valid for @xmath102 , which includes the physical limit for helioseismologically acceptable variations of the matter density . from figure 2 , we observe also that the maximal effect of the matter density and/or magnetic field fluctuation on the survival probability occurs if these magnetohydrodynamics fluctuations are found in the beginning of the convective zone where @xmath86 . this can be understood when we examine the behavior of the adiabaticity parameter for the spin - flavor conversion phenomenon . as we have already mentioned , this parameter increases with @xmath103 and approaches the unit for @xmath104 . in fact , it was pointed out previously ( see , for instance @xcite ) that this is a privileged situation for the neutrino conversion . note also that according to references @xcite , the required value of @xmath105 to find a solution to the solar neutrino anomaly is @xmath106 ev@xmath107 or smaller . if @xmath108 ev@xmath107 , neutrinos of @xmath80 mev in energy will experience their resonance around @xmath73 and therefore will be very sensitive to the maximal amplitude probability indicated in figure 2 . this is exactly the energy range probed by real time experiments , like as superkamiokande @xcite and sno @xcite , which are , therefore , very suitable to investigate the hypothesis of solar magnetohydrodynamics perturbations and their effects on solar neutrinos . after several hundred days of taking data , superkamiokande observations are , in principle , able to verify the existence of time fluctuation with period from @xmath79 to @xmath80 days . unfortunately , these data are not published nor available . nevertheless we can argue the potential of this experiment to perform such analysis . figure 3 shows the amplitude of the expected event rates in superkamiokande experiment when we assume that a magnetohydrodynamics fluctuation is localized around @xmath109 . we convulated the standard @xmath110-neutrino production spectrum @xcite , the @xmath111 scattering cross - section increasing linearly with energy and the experimental efficiency which is approximately 20% for neutrino energies around 5 mev up to a maximum of 70% for neutrino energies of 10 mev or larger . these event rates are given in solar neutrino units ( snu ) for several values of @xmath105 . we observe that for @xmath112 ev@xmath107 , the maximum event rate amplitude is obtained . i.e. , after having been precessed due to the existence of a sonic magnetohydrodynamical waves , the number solar neutrinos scattering with superkamiokande electrons fluctuates with amplitude around 0.003 - 0.004 snu , if these neutrinos present energies around 8 and 10 mev . this is approximately 10 - 15% of the total rate of scattering events in this energy range . note that if @xmath105 is different from the values shown in figure 3 , this means that neutrinos with energies around 5 - 15 mev will experience resonance in a position different from @xmath113 and , although not maximal , their counting rate will still fluctuate . magnetohydrodynamics predicts magnetic as well as matter density fluctuations in the sun . the purpose of this paper is to analyze the effects of possible solar matter density fluctuations on solar neutrino observations taking into account the stringent limits on these fluctuations coming from helioseismology . we verify also the potential of high statistics real time solar neutrino experiments , like as superkamiokande , to observe these effects . we showed that the survival probability fluctuations of active solar neutrinos due to the resonant neutrino spin - flavor precession can be of order 10% if amplitudes of the magnetic or matter density fluctuations @xmath114 or @xmath59 are of order of 5% ( 20% if @xmath11510% ) . therefore , in order to make evident the existence of these magnetohydrodynamics fluctuations in the sun , we have to look for solar neutrino data time fluctuations of this order in the time scale compatible with the periods predicted by the theory . from figure 1 , we see that relevant time scale can be of order @xmath79 to @xmath80 days . collecting @xmath116 events a day , superkamiokande presents a statistics error of approximately 10% in an interval of 5 days . furthermore , a fourier analysis of the experimental observations fitting a period of time of the order of the total period of superkamiokande observation ( of order of hundred days ) , could drastically reduce the involved errors and the amplitudes of order of 10% shown in figures 1 and 2 could be experimentally tested . there is no available published solar neutrino data to conduct a conclusive analysis of possible data fluctuation with period @xmath117 days . analysis of time fluctuations on solar neutrino data conducted up to now , privileged different periods : exact 24 hours as a result of msw effect inside the earth , six months to conciliate seasonal variation and neutrino vacuum oscillations or 11 years to verify solar magnetic activity effects on spin flip phenomenon . we conclude that superkamiokande is potentially interesting to investigating the effects on solar neutrino observations coming from magnetohydrodynamics fluctuations . the observation of these effects could be taken as an evidence of the resonant spin - flavor precession mechanism in the sun . * acknowledgements * : this work was supported by fundao de amparo pesquisa do estado de so paulo ( fapesp ) , programa de apoio a ncleos de excelncia ( pronex ) and conselho nacional de desenvolvimento cientfico e tecnolgico ( cnpq ) . a. cisneros , astrophys . space sci . 10 ( 1970 ) 87 ; m. voloshin , m.i . vysotsky and l. okun , sov . j. nucl . phys . 44 ( 1986 ) 440 ; j. schechter and j.w.f . valle , phys . d 24 ( 1981 ) 1883 ; e. akhmedov , sov . j. nucl . 48 ( 1988 ) 382 ; phys . b213 ( 1988 ) 64 ; c.s . lim and w.j . marciano , phys . d37 ( 1988 ) 1368 . k. lande ( homestake collaboration ) in _ neutrino 98 _ , proceedings of the xviii international conference on neutrino physics and astrophysics , takayama , japan , 49 june 1998 , edited by y. suzuki and y. totsuka , to be published in nucl . b ( proc . suppl . ) , scanned transparencies are available at url http://www-sk.icrr.u-tokyo.ac.jp/nu98/scan/index.html . krastev and a. yu . smirnov , phys . b 226 ( 1989 ) 341 ; mod . a 6 ( 1991 ) 1001 ; a. abada and s.t . petcov , phys . lett b 279 ( 1992 ) 153 ; f.n . loretti and a.b . balantekin , phys . d 50 ( 1994 ) 4762 . * figure 1 * : periods of the magnetohydrodynamics fluctuations for the continuum spectra when the magnetic profile shown in eq . ( [ prof1 ] ) when @xmath96 , @xmath87 , @xmath99 and exponential profile of eq . ( [ prof2 ] ) . we adopted the parameters @xmath121 g , @xmath122 g and @xmath123 [email protected] * figure 2 * : amplitude @xmath125p of the survival probability in function of @xmath83 for some values of @xmath126 . the maximum probability amplitude occurs around @xmath127 . the vertical lines correspond to a resonance coinciding with the singularity @xmath126 .	taking into account the stringent limits from helioseismology observations on possible matter density fluctuations described by magnetohydrodynamics theory , we find the corresponding time variations of solar neutrino survival probability due to the resonant spin - flavor precession phenomenon with amplitude of order @xmath0% ) . we discuss the physics potential of high statistics real time experiments , like as superkamiokande , to observe the effects of such magnetohydrodynamics fluctuations on their data . we conclude that these observations could be thought as a test of the resonant spin - flavor precession solution to the solar neutrino anomaly . .1 cm pacs numbers : 90.60.jw , 96.60.hv keywords : solar neutrinos , solar magnetic fields and waves
recent progress of optical clocks based on neutral atoms in optical lattices and on single ions in rf traps has led to fractional uncertainties of 10@xmath2 level@xcite . various implementation with different atomic species are being explored to reach and/or to reduce these uncertainties . it is pointed out that measurement of frequency ratio of different optical clocks is an essential step towards the redefinition of the second@xcite . the ratio measurement is of crucial importance also for testing the temporal variation of the fundamental physical constants@xcite . in the implementation with single ions , the optical frequencies have been reported so far with @xmath3yb@xmath1@xcite , @xmath4sr@xmath1@xcite , @xmath5ca@xmath1@xcite , @xmath6hg@xmath1@xcite , @xmath7al@xmath1@xcite , @xmath0in@xmath1@xcite . in@xmath1 is assumed to be a promising candidate of the original single - ion optical clock proposed by dehmelt@xcite , and frequency measurements with fractional uncertainties in 10@xmath8 level have been reported previously@xcite . however , there has been no update since 2007 in contrast to other ions reporting repeated improvements in the level below 10@xmath9 . frequency discrepancy of about 1 khz by two research groups remains also unresolved@xcite . recent theoretical prediction of small black body radiation ( bbr ) shift revives the advantage of in@xmath1 as an ion of the iiia group@xcite . frequency measurement of in@xmath1 with more precision than those of previous reports will make in@xmath1 a useful candidate of the frequency ratio measurement . it should be also noted that in@xmath1 is a promising candidate of the multi - ion optical clock with enhanced stability @xcite . in this paper we demonstrate spectroscopy of the in@xmath1 clock transition in a new approach based on sympathetic cooling , and report its frequency measurement using a local frequency reference , utc(nict ) ( realization of universal coordinated time at nict ) and a sr optical lattice clock . optical clock . ( a ) relevant energy levels and transitions in in@xmath1 , and ( b ) those of ca@xmath1 . ( c ) schematic drawing of the ion chain configurations for our new approach . ] the ion - based optical clock requires methods of cooling and of state detection in addition to the access to the clock transition . the @xmath10s@xmath11-@xmath10p@xmath12 transitions to provide the two methods in the iiia group ions are located in the vacuum ultraviolet range , and technical difficulties to prepare the single - mode light sources prevent the implementation of the optical clocks . this difficulty has been resolved by quantum logic spectroscopy ( qls ) of al@xmath1 , in which another ion is used for the cooling and the detection@xcite . the disadvantage of the qls is high technical demands for the cooling to the ground state of motion , and repeated detection for fidelity state readout . in@xmath1 has an exceptional feature as an iiia group ion that the @xmath10s@xmath11-@xmath13p@xmath12 transition can supply direct state detection and laser cooling owing to its relatively large linewidth of 360 khz as shown in fig.[concept](a ) . this feature was employed to direct quantum state detection@xcite , and to bichromatic laser cooling@xcite . the bichromatic laser cooling requires two laser frequencies to stabilize the cooling . the previous implementations of the in@xmath1 optical clock rely on these methods@xcite . we demonstrate a new approach , which combines the concepts of sympathetic cooling used in qls , and of the direct state detection in the previous in@xmath1 optical clocks@xcite . choice of ca@xmath1 as shown in fig . [ concept](b ) provides an easier access to cooling using compact diode - laser - based systems @xcite . when doppler cooling is applied to a ca@xmath1 , a sympathetically cooled in@xmath1 can be prepared in a simpler way than by bichromatic cooling . the direct quantum state detection of the in@xmath1 allows an easier implementation of the optical clock than by qls . although the frequency inaccuracy achieved ultimately with this approach is limited by the doppler cooling limit temperature of ca@xmath1 , the sideband cooling to the ground state using the @xmath14s@xmath15-@xmath14d@xmath16 clock transition could be used to reduced the inaccuracy by time dilation@xcite . the ca@xmath1 can be used also as a probe for the trapping and surrounding fields . the use of the @xmath14s@xmath15-@xmath14p@xmath15 transition with a larger scattering rate provides a fast evaluation of micromotion and the magnetic field . more precise field evaluation might be implemented by using the clock transition . the minimum configuration for the sympathetic cooling is to add one ca@xmath1 . however , asymmetric application of the trapping field is necessary to locate the in@xmath1 at the same site in this configuration . it is essential to place the ion at the same site to avoid location - dependent systematic shifts . use of two ca@xmath1 ions instead of one enables a symmetric configuration depicted as ( i ) in fig.[concept](c ) . this configuration is very convenient for the clock transition spectroscopy , because it is relatively easily recovered whenever it changes into the configuration ( ii ) . the recovery was demonstrated with amplitude modulation on the cooling laser to destabilize the configuration ( ii ) only@xcite . in the present study we use an even simpler method to reduce the trapping rf voltage until the heavier ion comes back to the center . the configuration ( ii ) with a ca@xmath1 at the center might be used for probing the electromagnetic field at the center for more precise systematic shift evaluation in future . for these practical advantages we use an ion chain consisting of one in@xmath1 and two ca@xmath1 . we demonstrate spectroscopy of the clock transition using such an ion chain for the first time . the overview of the experimental setup is shown in fig . [ setup ] . a linear rf trap is fabricated with becu plates with gold plating on both surfaces to avoid charging due to the deep uv beams . two dc electrodes are segmented in three parts to supply confinement along the trap axis , while two rf electrodes have short slits as shown in fig . [ concept](c ) . the distance from the trap center to the electrode edge is 0.5 mm , and the width of the center dc electrode is 3 mm . the trap electrodes are fixed to insulators made of glass ceramic , and then attached to a support consisting of machined becu blocks installed in a octagon - shaped vacuum chamber made of non - magnetic stainless . the chamber with uv - grade sapphire windows is evacuated to a pressure below 10@xmath17 pa . the trap is driven with an rf frequency of 27 mhz and a dc with 500 v , which gives ca@xmath1 secular frequency of 4.0 mhz and 0.5 mhz for the radial and the axial directions respectively . these voltages give radial secular frequency of 1.4 mhz for in@xmath1 and in - phase mode frequency along the axial direction of 0.4 mhz for the configuration ( i ) , which bring the in@xmath1 in the lamb - dicke regime . the light source system for ca@xmath1 consists of a 397-nm and a 866-nm extended - cavity diode lasers ( ecdls ) for laser cooling , and of a 423-nm ecdl , and a 380-nm diode laser for resonant photo ionization . spectral filtering with a filter cavity or a grating , and frequency stabilization using transfer cavity referenced to a frequency - stabilized he - ne laser are implemented to the 397-nm and 866-nm lasers @xcite . the light source system for in@xmath1 consists of a 411-nm ecdl for resonant photo - ionizaiton , a 230-nm source for detection and a 237-nm clock laser . the 230-nm radiation is generated by two - stage frequency doubling of an amplified 922-nm ecdl . the master laser is locked to a ule ( ultra low expansion ) cavity with a finesse of about 10,000 . about 200 mw of output from the amplifier is frequency - doubled twice , first in a periodically - poled lithium niobate ( ppln ) waveguide to 60 mw at 461 nm , then in a beta - barium borate ( bbo ) crystal placed in an enhancement cavity , to the final output of 1 mw at 230 nm . the 237 nm radiation for the clock transition is generated by two - stage frequency doubling of an amplified diode laser at 946 nm in a similar way , except the bulk ppln crystal in an enhancement cavity in the first stage instead of a waveguide . available output power at 237 nm is 15 mw . we stabilize the fundamental frequency of 946 nm to a high finesses ule cavity ( finesse 320,000 ) , and control the ule cavity at the measured zero - crossing temperature of 32.8 @xmath18c . the linear drift of the cavity resonance is substantially eliminated by an aom , to which counter - drifting frequency is introduced from a dds - based oscillator . this drift compensation suppressed the long - term drift below 0.05 hz / s . simultaneous detection of the two ion species is performed with an image - intensified ccd ( iccd ) camera for ca@xmath1 and a solar - blind photo - multiplier tube ( pmt ) with a quantum efficiency of 30% for in@xmath1 . the separate imaging of two wavelengths resolves the large chromatic aberration for 397 nm and 230 nm of the imaging lens with focal length of 50 mm and an effective aperture of 27 mm . the configuration of the ions is continuously monitored with the iccd camera , and is kept to the configuration ( i ) . whenever the chain changes to the configuration ( ii ) , the measurement is interrupted and the rf potential is lowered and then heightened for reordering the ions to ( i ) . this simple method of ion reordering relies on mass dependence of the pseudo potential , and is effective only when the heavier ion is placed to the center . details of the method will be reported elsewhere . spectroscopic data with the configuration ( ii ) are not recorded in the present experiments . the experimental procedure starts with optimization of the micromotion compensation voltages and the magnetic field using a single ca@xmath1 followed by loading of another ca@xmath1 and an in@xmath1 , then the spectra are recorded by scanning the clock laser frequency by acousto - optic modulator ( aom ) . the frequency of the clock laser is measured using an optical frequency comb during the spectrum measurement , and the clock transition frequency is determined after each measurement session by taking the average of the scan data . the stabilization of the clock laser to the transition was not yet implemented in the present study as the spectroscopy signal often falls unavailable due to the recurrent changes of ion configuration . the ca@xmath1 is loaded by resonant photo - ionization of thermal ca beam from an oven installed beneath the trap using the 423-nm and 380-nm lasers@xcite . the ion is laser - cooled with the 397-nm and 866-nm ecdls . the number of ca@xmath1 is adjusted to one for the optimization of the micromotion compensation voltages and the magnetic field . micromotion compensation for three directions is performed with the ca@xmath1 using the standard procedure based on the fluorescence - rf correlation technique @xcite . the magnetic field at the trap center is controlled by three pairs of coils orthogonally attached to the vacuum chamber . the coil currents to null the magnetic field are searched for by using the dark states formed in the @xmath14d@xmath19 state of ca@xmath1@xcite . by iterative minimization of the ca@xmath1 fluorescence with currents of each coil pair , a set of currents to minimize the magnetic field is found@xcite . after the optimization of the fields , two ca@xmath1 are loaded in the similar way . the number of ca@xmath1 is adjusted to two by controlling the trap rf voltage@xcite . then an in@xmath1 is loaded by resonant photo - ionization with the 411-nm ecdl . the in@xmath1 is sympathetically cooled with the ca@xmath1 . the number of in@xmath1 is identified as dark sites in the iccd image , and is adjusted by ejecting extra in@xmath1 ions by controlling the trap rf voltage@xcite . by measuring the in - phase mode frequency of the chain , the final dark site is identified as in@xmath1 with the mass number of 115 . the quantum state detection of the in@xmath1 is performed using the pmt . only when the in@xmath1 is in the @xmath10s@xmath11 state the ion emits photons at 230 nm and gives detection of about 15 photons per 60 ms gate time with approximately 10@xmath20w of the 230-nm light . this count clearly distinguishes that from scattered light and the dark count of 3 , with a discrimination fidelity of about 90% . spectroscopy of the clock transition is performed by recording the excitation probability as a function of the aom frequency ( @xmath21 ) for scanning the clock laser frequency . in order to avoid shifts due to the lasers other than the clock laser , the ca@xmath1 cooling lasers as well as the 230-nm in@xmath1 detection beam are blocked by mechanical shutters during the clock pulse irradiation periods . 200 trials are made for a single @xmath21 to obtain the excitation probability in the following procedure . the first step is to prepare the in@xmath1 in the @xmath10s@xmath11 state in the configuration ( i ) . this is confirmed by the image on the iccd and the count of the pmt . then all laser beams are blocked by the shutters , and the 237-nm clock pulse for a duration of 20 ms illuminates the in@xmath1 . the clock laser power of about 1.1 @xmath20w before the shutter and the estimated spot size diameter 100 @xmath20 m at the trap center gives similar intensity as in the previous reports@xcite . after the clock pulse irradiation other lasers are turned on , and the quantum state of the in@xmath1 is determined by the 230-nm fluorescence for the detection period of 60 ms . the single cycle takes 100 ms on average including dead times of 0.8 ms to wait for the mechanical shutters to be settled and process time of about 20 ms by pc . while the shutter timings are precisely generated by a delay generator , the process time by pc fluctuates due to the non - real - time nature of the operating system . repeating this procedure typically 200 times the excitation probability for a single aom frequency @xmath22 is obtained . by scanning @xmath21 with a step of 80 hz at 237 nm the spectrum of the in@xmath1 clock transition is obtained . a @xmath23/4 plate attached to a motorized rotation mount switches the polarizations of 230 nm and 237 nm between the settings of ( @xmath24 + , @xmath24- ) and ( @xmath24- , @xmath24 + ) . the spectrum of the transition from @xmath25 to @xmath26 is first recorded in the former setting , then that from @xmath27 to @xmath28 in the latter setting . the transition centers ( @xmath29 ) are derived by curve fittings to lorentzian . the average of @xmath30 and @xmath31 gives the clock transition frequency @xmath32 with the 1st order zeeman shift canceled @xcite . in the following the set of the two records is defined as one measurement . a snapshot of the measurement with a magnetic field along @xmath33-axis ( @xmath34 ) of about 8.8 @xmath20 t is shown in fig . [ f - spectra ] with lorenz curves for the least - square fitting . the linewidth obtained by the fitting ranges from 350 to 700 hz . we confirmed that zeeman - shift coefficient of @xmath30 and @xmath31 agrees well with the reported values of the two transitions @xcite . clock transition spectra for the two settings of the laser polarizations . the bars represent measured excitation probability , while the curves are lorenz curves for the least square fitting . ( b ) observed dependence of the spectrum center on the magnetic field along z - direction @xmath34 . the red filled circles shows those for @xmath35 clock excitation , while the blue empty circles represents those with @xmath36 clock excitation . the lines are linear fit to the data . ] frequency determination and uncertainty evaluation have been performed as follows using the data taken in three measurement sessions , which include 23 , 6 , 7 measurements , respectively . during all the sessions the clock laser frequency at 946 nm was measured with an er - fiber - based optical frequency comb in the methods shown in fig . [ f - scheme].the @xmath32 values of all measurements determined by the frequency comb are plotted in fig . [ f - frequency ] , in which error bars correspond one standard deviation of the fitting . the mean value and the statistical uncertainty of each session are determined by the @xmath32 values by using the fitting errors as the weight . the comb was referenced to utc(nict ) in session a and b , while it was referenced to a @xmath37sr optical lattice clock in session c @xcite . for the measurements in session a and b the frequency of one component of the frequency comb ( @xmath38 ) was phase - locked to the the clock laser at 946 nm ( @xmath39 ) , and we recorded the phase of the down - converted microwave relative to utc(nict ) in every second , by which we obtained the mean frequency of @xmath39 of one measurement . utc(nict ) is calibrated with reference to the si second by so - called tai - link ( international atomic time - link).the fractional uncertainty of the link between utc(nict ) and the si second amounts to @xmath40 considering the uncertainty published in circular t as well as dead time uncertainty of the utc(nict ) and utc - si second@xcite . this corresponds to 3.3 hz for the @xmath41 clock transition at 237 nm . in session c , @xmath38 was phase - locked to the clock laser of the sr optical lattice clock ( @xmath42 ) , and the beat frequency between the 946 nm clock laser and the nearest comb component was recorded by a conventional frequency counter ( agilent 53132a ) . the absolute frequency of the sr lattice clock was recently evaluated with fractional uncertainty of @xmath43@xcite . in session a , @xmath34 was swept from -14 @xmath20 t to 14 @xmath20 t in order to confirm that the transitions of proper zeeman sublevels were excited , and to evaluate the uncertainty of the residual 1st order zeeman shift . measurements with residual magnetic field smaller than 5@xmath20 t were performed in other sessions . the last 6 digits of the averaged frequency of each session are summarized in table . [ result ] with errors including statistical error and the uncertainty of the link to the si second . tiny difference of the values in session a and b shows that the simple cancellation method of the 1st order zeeman shift is effective to this degree . the agreement between session b and c implies that no bias is present in the frequency link to the si second@xcite . the weighted average of the sessions gives 1 267 402 452 901 060.4 ( 5.0 ) , which does nt include systematic uncertainties of the shift in the in@xmath1 transition associated with fields and motion . in general , the statistical uncertainty can be reduced by increasing the number of measurements . however , we found that the further reduction of the statistical uncertainty by more measurement sessions would be ineffective , because the systematic uncertainty discussed in the following limits the overall frequency uncertainty in the present study . in three sessions and of their weighed averaged value denoted as the gray line . the two gray dotted lines represent a range of the standard error . systematic shifts including gravitational shift are not taken into account in this figure . ] .measurement result of the three measurement sessions . the error includes statistical and uncertainty of the link to the si second . systematic errors of the ion are not included . n _ : number of measurements , @xmath44 : magnetic field strength . [ cols="^,^,^,^,^",options="header " , ] [ systematics ] table . [ systematics ] summarizes the shifts and uncertainties of the in@xmath1 transition . the largest systematic error in the present study comes from the 1st order zeeman shift . imperfect polarization of the 230 nm and 237 nm beams , and residual magnetic field along _ x _ and _ y _ directions allow excitation of transitions between unwanted zeeman sublevels . this gives asymmetry to the two spectrum centers @xmath30 , @xmath31 defining one measurement , leading to a frequency shift . this shift was estimated by checking the dependence of calculated center frequency @xmath32 on the magnetic field @xmath34@xcite . from the data in session a we estimate the residual shift dependent linearly on @xmath34 with -912 hz / mt . the shift should be canceled at the zero magnetic field , but the measurements at low magnetic fields in session b , c show spectrum linewidths larger than 350 hz , which suggests that possible residual magnetic field in _ x_- and _ y_-directions is not smaller than 5 @xmath20 t . we assigned @xmath454.6 hz calculated for the residual field of 5 @xmath20 t as the uncertainty for the 1st order zeeman shift . the second - order zeeman shift of in@xmath1 is known to be much smaller than those of other ions , and is given as 4.1@xmath46 hz / mt@xmath14@xcite , in which the square of the magnetic field @xmath47 includes contributions from the residual magnetic field and the alternating magnetic field due to unbalanced currents of the ion trap rf field . both contributions are estimated as much smaller than 0.1 hz . the ac - stark shift due to the clock laser beam is estimated by calculating the shifts of the @xmath10s@xmath11 and the @xmath13p@xmath11 states due to coupling to the energy levels listed in ref . we added uncertainty of 0.1 hz as the upper limit obtained by the calculation . the contribution of the other lasers are not taken into account , because they are completely shut off during the clock laser irradiation periods by mechanical shutters . the bbr shift of in@xmath1 is recognized as smaller than those of other optical clock candidates except al@xmath1 , and is estimated as 17 mhz at 300 k@xcite . the largest contribution in the temperature uncertainty comes from the rf electrodes continuously exposed to the rf drive . if we assume the temperature of the electrodes rises to as high as 200 @xmath18c , and the in@xmath1 ion sees the electrodes with 4@xmath48 solid angle , the bbr shift would be 0.1 hz . considering the actual solid angle much smaller and the electrode temperature not higher than this assumption , we estimate the bbr shift uncertainty smaller than 0.1 hz . the frequency shift due to the time dilation comes from the residual micromotion and secular motion of the in@xmath1 ion . the uncertainty by micromotion was estimated from the rf correlation signal using the single ca@xmath1 prior and after the frequency measurements in the standard method@xcite , and is estimated to be 0.4 hz . the rather large uncertainty comes from the low sensitivity of our rf correlation measurement of the radial direction ( @xmath49 in fig.[concept](c ) ) limited by small angle of one of the 397-nm laser beams , which could be improved in future by modifying the beam irradiation geometry . strict evaluation of the uncertainty due to residual secular motion by sideband spectra was not performed in the present study due to unavailability of a vacuum chamber port to probe the secular motion of in@xmath1 along all the directions . instead we estimated the worst case , in which vibrational wavepacket size of the ca@xmath1 is as large as the image on the iccd camera . the 3@xmath20 m rms size corresponds to the maximum rms velocity 5.7 m / s of in@xmath1 assuming the in - phase mode of the ion chain@xcite . this leads to the uncertainty of 0.7 hz . the gravitational shift is well characterized in other experiments done in the same floor of the laboratory building @xcite , and is 10.5 hz for in@xmath1 . the systematic shifts associated with the clock frequency scan is evaluated to be sufficiently small compared to other shifts . the clock transition frequency is determined to be 1 267 402 452 901 049.9 ( 6.9 ) hz from the measurements and the above discussions . figure [ f - comparison ] represents the comparison of the measured frequency with reported values . in the previous studies a methane - stabilized he - ne laser or a commercial cesium beam clock both calibrated at remote sites was used as a frequency reference @xcite . in contrast our measurement has been performed using two different frequency standards located in the same site , and , therefore , it presumably provides more reliable measurements . in conclusion we have performed a measurement of the clock transition frequency of an in@xmath1 with a fractional uncertainty of 5@xmath5010@xmath51 . it is the first measurement of the in@xmath1 transition by new approach using sympathetic cooling in a linear trap . our measurement determines the absolute clock transition frequency with the smallest uncertainty ever reported . the optical clock operation will be implemented by locking the clock laser to the transition , when the instability of the ion - chain arrangements is reduced by further studies . the advantages of in@xmath1 represented by the absence of the quadrupole shift and by the immunity to the bbr shift will be fully exploited in our future work including extension to the multi - ion optical clock . we thank h. hachisu for his contribution in the operation of the sr lattice clock , and r. locke for his contribution in the early stage of the project . the stable optical cavity for 922 nm laser was assembled by k. etoh . we are also grateful to m. hosokawa and y. hanado for fruitful discussions . the project was achieved with the continuous technical help provided by h. ishijima , m. mizuno . this work has been supported by jsps and daad under the japan - germany research cooperative program . r. m. godun , p. b. r. nisbet - jones , j. m. jones , s. a. king , l. a. m. johnson , h. s. margolis , k. szymaniec , s. n. lea , k. bongs , and p. gill , frequency ratio of two optical clock transitions in @xmath3yb@xmath1 and constraints on the time variation of fundamental constants , phys . lett . * 113 * , 210801 ( 2014 ) . p. dub , a. a. madej , z. zhou , and j. e. bernard , evaluation of systematic shifts of the @xmath4sr@xmath1 single - ion optical frequency standard at the 10@xmath52 level , phys . a * 87 * , 023806 ( 2013 ) . h. s. margolis , g. p. barwood , g. huang , h. a. klein , s. n. lea , k. szymaniec , and p. gill , hertz - level measurement of the optical clock frequency in a single @xmath4sr@xmath1 ion , science * 306 * , 13551358 ( 2004 ) . m. chwalla , j. benhelm , k. kim , g. kirchmair , t. monz , m. riebe , p. schindler , a. s. villar , w. hnsel , c. f. roos , r. blatt , m. abgrall , g. santarelli , g. d. rovera , and ph . laurent , absolute frequency measurement of the @xmath5ca@xmath1 4s@xmath14s@xmath15 - 3d@xmath14d@xmath16 clock transition , phys . lett . * 102 * , 023002 ( 2009 ) . k. matsubara , h. hachisu , y. li , s. nagano , c. locke , a. nogami , m. kajita , k. hayasaka , t. ido , and m. hosokawa , direct comparison of a ca@xmath1 single - ion clock against a sr lattice clock to verify the absolute frequency measurement , opt . express * 20(20 ) , 22034 - 22041 ( 2012 ) . y. huang , j. cao , p. liu , k. liang , b. ou , h. guan , x. huang , t. li , and k. gao , frequency comparison of two @xmath5ca@xmath1 optical clocks with an uncertainty at the 10@xmath52 level , phys . 116 * , 013001 ( 2016 ) . t. rosenband , d. b. hume , p. o. schmidt , c. w. chou , a. brusch , l. lorini , w. h. oskay , r. e. drullinger , t. m. fortier , j. e. stalnaker , s. a. diddams , w. c. swann , n. r. newbury , w. m. itano , d. j. wineland , and j. c. bergquist , frequency ratio of al@xmath1 and hg@xmath1 single - ion optical clocks ; metrology at the 17th decimal place , science * 319 * , 1808 - 1812 ( 2008 ) . chen , s. m. brewer , c. w. chou , d. j. wineland , d. r. leibrandt , and d. b. hume , sympathetic ground state cooling and time - dilation shifts in an @xmath7al@xmath1 optical clock , phys . lett . * 118 * , 053002 ( 2017 ) . j. von zanthier , th . becker , m. eichenseer , a. yu . nevsky , ch . schwedes , e. peik , h. walther , r. holzwarth , j. reichert , th . udem , t.w . hnsch , p.v . pokasov , m.n . skvortsov , and s.n . bagayev , absolute frequency measurement of the in@xmath1 clock transition with a mode - locked laser , opt . lett . , * 25(23 ) , 17291731 ( 2000 ) . y. h. wang , r. dumke , t. liu , a. stejskal , y. n. zhao , j. zhang , z. h. lu , l. j. wang , th . becker , and h. walther , absolute frequency measurement of the @xmath0in@xmath1 5s@xmath14 @xmath10s@xmath11 - 5s5p@xmath13p@xmath11 narrowline transition , opt . commun . 273 * , 526 - 531 ( 2007 ) . s. uetake , k. matsubara , h. ito , k. hayasaka , and m. hosokawa , frequency stability measurement of a transfer - cavity - stabilized diode laser by using an optical frequency comb , appl . b * 97 * , 413 - 419 ( 2009 ) . g. p. barwood , p. gill , g. huang , h. a. klein , and w. r. c. rowley , sub - khz clock transition linewidths in a cold trapped @xmath4sr ion in low magnetic fields using 1092-nm polarisation switching , opt . commun . * 151 * , 5055 ( 1998 ) . a. d. ludlow , m. m. boyd , t. zelevinsky , s. m. foreman , s. blatt , m. notcutt , t. ido , and j. ye , systematic study of the @xmath37sr clock transition in an optical lattice , phys . 96 * , 033003 ( 2006 ) .	we report frequency measurement of the clock transition in an @xmath0in@xmath1 ion sympathetically - cooled with ca@xmath1 ions in a linear rf trap . the ca@xmath1 ions are used as a probe of the external electromagnetic field and as the coolant for preparing the cold in@xmath1 . the frequency is determined to be 1 267 402 452 901 049.9 ( 6.9 ) hz by averaging 36 measurements using an optical frequency comb referenced to the frequency standards located in the same site .
a quasi - exactly solvable quantum - mechanical quartic oscillator was introduced in @xcite and ( in its restricted form ) is a schrdinger - type eigenvalue problem of the form @xmath0 with the boundary conditions @xmath1 as @xmath2 where @xmath3 and @xmath4 are parameters of the problem.[multiblock footnote omitted ] with these boundary conditions , real @xmath5 and @xmath4 , is not hermitian but @xmath6-symmetric , see @xcite , @xcite and @xcite . when @xmath7 is a positive integer , then @xmath8 maps the space of quasi - polynomials @xmath9 to itself where @xmath10 is a polynomial of degree at most @xmath11 and @xmath12 . the restriction of @xmath13 to the latter space is a finite - dimensional linear operator whose spectrum and eigenfunctions can be found explicitly . this part of the spectrum and eigenfunctions of is usually referred to as _ quasi - exactly solvable_. with a slight abuse of notation , one can easily show that polynomial factors in the quasi - exactly solutions of are exactly polynomial solutions of the degenerate heun equation @xmath14 where @xmath3 has the same meaning as above and @xmath15 are spectral variables . obviously , if equation has a polynomial solution of degree @xmath16 then @xmath17 . additionally , to get a polynomial solution of of degree @xmath11 , the remaining spectral variable @xmath18 should be chosen as an eigenvalue of the operator @xmath19 acting on the linear space of polynomials of degree at most @xmath20 in the standard monomial basis @xmath21 the operator @xmath22 acts by the @xmath23-diagonal @xmath24 square matrix @xmath25 given by : @xmath26 we will call the bivariate polynomial @xmath27 the _ @xmath11-th spectral polynomial of _ . the total degree of @xmath28 equals @xmath29 which is also its degree with respect to the variable @xmath18 . ( the degree of the variable @xmath5 in @xmath28 equals @xmath30 $ ] . ) observe that if @xmath5 is a real number then @xmath28 is a real polynomial in @xmath18 and therefore the spectrum is symmetric with respect to the real axis . our main goal is to study the behavior of the spectrum of @xmath31 when @xmath32 for two regimes of @xmath5 , namely , when @xmath33 or when @xmath34 . in other words , we consider solutions of @xmath35 with respect to @xmath18 in the above two situations . observe that if @xmath36 does not exist one can not expect any reasonable asymptotic behavior of the spectrum . _ in case of a non - degenerate heun equation detailed study of similar asymptotics was carried out in @xcite , @xcite by the present authors jointly with k.takemura . our mathematically rigorous result is as follows . [ conj : main ] ( i ) if @xmath33 , the maximal absolute value @xmath37 of the eigenvalues of @xmath38 i.e. , the maximal absolute value among the roots of @xmath39 grows as @xmath40 . \(ii ) if @xmath33 , the sequence @xmath41 of root - counting measures for the sequence @xmath42 , i.e. , for the sequence of spectra of @xmath43 rescaled so that each eigenvalue of @xmath31 is divided by @xmath44 , weakly converges to the measure @xmath45 supported on the union of three straight intervals connecting the origin with three cubic roots of unity multiplied by @xmath46 , see figure [ fig1 ] . more information about @xmath45 can be found in 2 . scaled by @xmath44 for n=200 . ] our main result on a physics level of rigor is as follows . consider the family @xmath47 of quadratic equations in @xmath48 depending on parameters @xmath5 and @xmath49 . [ quart ] if @xmath34 , the sequence @xmath41 of root - counting measures for @xmath50 weakly converges to a compactly supported probability measure @xmath51 . the support of @xmath51 consists of all values of @xmath52 for which there exists a compactly supported in the @xmath53-plane probability measure @xmath54 whose cauchy transform @xmath55 satisfies almost everywhere in the @xmath53-plane . recall that for a ( complex - valued ) measure @xmath56 compactly supported in @xmath57 , its logarithmic potential is defined as @xmath58 and its cauchy transform is defined as @xmath59 the existence of a signed ( not necessarily positive ) measure @xmath56 whose cauchy transform @xmath60 satisfies almost everywhere in @xmath57 is closely related to the properties of the family @xmath61 of quadratic differentials depending on @xmath5 and @xmath49 . more information about this connection can be found in 3 . for basic information on quadratic differentials and its horizontal trajectories consult @xcite . proposition [ quart ] reminds of the main result of @xcite , where we considered the so - called homogenized spectral problem . in this case all terms of the heun differential operator in question are important for the root asymptotics of sequences of eigenpolynomials . the validity of proposition [ quart ] was also verified numerically . finally our main conjecture supported by extensive numerical experiments is as follows . for a given positive integer @xmath11 , denote by @xmath62 the set of all branching points of the projection of the algebraic curve @xmath63 to the @xmath5-axis . in other words , @xmath62 is the set of all values of the complex parameter @xmath5 for which @xmath28 has a multiple root in @xmath18 . equivalently , @xmath62 is the zero locus of the univariate discriminant polynomial @xmath64 which is the resultant of @xmath28 with @xmath65 . one can show that the degree of the latter polynomial equals @xmath66 . yablonskii - vorobev polynomials are defined as follows , see @xcite @xcite . set @xmath67 . for @xmath68 set @xmath69 although the latter expression apriory determines a rational function , @xmath70 is in fact a polynomial of degree @xmath71 , see e.g. @xcite . the importance of yablonskii - vorobev polynomials is explained by the fact that all rational solutions of the second painlev equation @xmath72 are presented in the form @xmath73\right\},\ ; u(t,0)=0,\ ; u(t ;- n):=-u(t;n).\ ] ] denote by @xmath74 the zero loci of @xmath75 . one can show that the maximal absolute value of points in @xmath62 grows as @xmath76 while the maximal absolute value of points in @xmath77 grows as @xmath78 . [ conj : yab ] ( i ) after division by @xmath76 and @xmath78 respectively , the sequence @xmath79 tends to the sequence @xmath74 . in particular , the limiting triangular domain @xmath80 covered by @xmath79 and @xmath74 when @xmath32 is the same . \(ii ) the boundary of @xmath80 consists of all values of @xmath81 for which the support of measure @xmath51 consists of a single arc with a singular point on it . for more details on conjecture [ conj : yab ] see 4 and for its illustration see fig . [ bbigtriangle ] . of the branching points for @xmath82 together with the zero locus @xmath83 after scaling . ( observe the surprising closeness of the scaled @xmath84 and @xmath83 . their distinction is hardly visible by a naked eye ! ) ] _ acknowledgements . _ the authors are sincerely grateful to m. duits , a. eremenko , a. kuijlaars , and p. larsson for many illuminating discussions of the topic . we start with the spectral asymptotics in the basic case @xmath85 . [ conj:2 ] ( i ) the spectrum of @xmath86 i.e. , the zero locus of @xmath87 , is invariant under the rotation by @xmath88 around the origin , i.e. , it has a @xmath89-symmetry . \(ii ) the sequence @xmath90 splits into the following three subsequences : 1 . if @xmath91 , the polynomial @xmath90 contains only the third powers of @xmath18 and denoting @xmath92 we have @xmath93 ; 2 . if @xmath94 , the polynomial @xmath90 can be factorized as @xmath95 ; 3 . if @xmath96 , the polynomial @xmath90 can be factorized as @xmath97 . \(iii ) all three polynomials @xmath98 have the same degree @xmath99 all their roots are negative and simple which implies that the spectrum is located on the union of three rays through the origin as illustrated in figure [ fig1 ] . to prove proposition [ conj:2 ] following @xcite , we need to introduce a double indexed polynomial sequence containing our original sequence @xmath100 . the most natural way to do this is to consider the principal minors of @xmath101 given by : @xmath102 namely , denote by @xmath103 the @xmath104-th principal minor of . obviously , @xmath105 . since the latter matrix is @xmath23-diagonal with one subdiagonal and two superdiagonals , then , by a general result of @xcite , its principal minors satisfy a linear recurrence relation of order at most @xmath106 ( i.e. , a @xmath23-term relation ) . simple explicit calculation gives @xmath107 where @xmath104 runs from @xmath108 to @xmath109 with the standard boundary conditions @xmath110 recurrence has variable coefficients depending both on @xmath104 and @xmath11 . another form of which looks more promising is as follows . to simplify the signs , let us consider the characteristic polynomials of the principal minors of @xmath111 ( instead of @xmath112 as above ) . with a slight abuse of notation , the recurrence relation looks as : @xmath113 items ( i ) and ( ii ) follow immediately from the next statement . [ lm:1 ] denote @xmath114 , @xmath115 , @xmath116 , where @xmath117 and @xmath118 are monic polynomials of degree @xmath119 . then in terms of @xmath118 recurrence looks as : @xmath120 with the initial conditions @xmath121 . here @xmath119 runs from @xmath108 to @xmath122 $ ] . simple algebra . to prove item ( iii ) of proposition [ conj:2 ] , we will use notation of lemma 4 . hence @xmath123 , @xmath124 , @xmath125 correspond to @xmath126 @xmath127 and @xmath128 respectively . our first step is to prove that each polynomial in the recurrences has positive coefficients and real roots . moreover the roots of the polynomials appearing in the same recurrence are interlacing . note that , in our case , positive coefficients imply that all roots are negative . we will use induction on @xmath119 . as the base of induction we check case @xmath129 since case @xmath130 is trivial . elementary calculations give @xmath131 and @xmath132 , which proves that the roots of each one of them are negative and are interlacing for all integer @xmath133 . hence , since @xmath134 , then using lemma 1.10 of @xcite or conducting elementary calculations , we can conclude that @xmath135 has real roots . using similar elementary calculations we obtain similar results for the pairs @xmath136 and @xmath137 . assume now that our hypothesis holds for a given positive integer @xmath119 . note that , in recurrence for @xmath119 , degrees of the polynomials differ by one . this indicates that the largest root belongs to the polynomial with larger degree . furthermore by corollary 1.30 of @xcite , we can derive the following results : @xmath138 @xmath139 @xmath140 here the arrow @xmath141 indicates that the corresponding pair of polynomials have interlacing roots with the largest root belonging to the polynomial at which the arrow points . 1 . consider the reccurence @xmath142 we can rewrite @xmath143 as @xmath144 observe that @xmath145 by induction hypothesis and corollary 1.30 in @xcite . using ( 2.2 ) , we can conclude that @xmath146 since all roots of @xmath147 are negative . hence @xmath148 , by lemma 1.31 in @xcite . therefore @xmath149 has real roots , by lemma 1.10 in @xcite . furthermore , since @xmath147 and @xmath150 have positive coefficients , so does @xmath149 , hence it has negative roots . 2 . consider the recurrence @xmath151 one can rewrite @xmath149 as in part @xmath152 . by ( 2.2 ) , @xmath153 and by ( 2.3 ) , @xmath154 since @xmath150 and @xmath155 only have negative roots . proof is concluded by the same argument as in part @xmath152 . 3 . consider the recurrence @xmath156 one can rewrite @xmath157 as in part @xmath158 . by ( 2.3 ) , @xmath160 by part @xmath152 , corollary 1.30 in @xcite and the fact that both @xmath149 and @xmath150 have only negative roots . proof is concluded by the same argument as in parts @xmath152 and @xmath158 . second step is to prove that all roots of the considered polynomials are simple . assume that @xmath147 has a double root @xmath10 . then by corollary 1.30 in @xcite , @xmath10 should also be a root of @xmath161 and @xmath162 . hence when we conduct our induction on @xmath119 and assume that @xmath161 and @xmath162 have only simple roots , we get a contradiction since @xmath147 is a linear combination of those polynomials . note that , for @xmath129 , we only have simple zeros . the same argument applies to @xmath155 and @xmath150 . in order to apply the approach of @xcite , we need the recurrence coefficients to stabilize when @xmath163 , for any fixed @xmath164 $ ] . this stabilization would hold if instead of one considers @xmath165 where @xmath166 . namely , when @xmath167 the latter relation tends to following relation with constant coefficients : @xmath168 observe that is the recurrence for the principal minors of the matrix @xmath169 given by : @xmath170 in other words , is satisfied by the characteristic polynomials of the principal minors of @xmath171 . following @xcite , we conclude that the cauchy transform ( considered in the complement @xmath172 where @xmath173 is an appropriate compact subset ) of the asymptotic distribution of the sequence @xmath174 is obtained by averaging the cauchy transforms of the asymptotic distributions of over @xmath164 $ ] . in fact , in the case under consideration even the density of the former distribution can be obtained by averaging the densities of the latter family of distributions . recurrence is similar to the one considered in the last section of @xcite and has very nice asymptotic distribution of its roots , see figure [ figtau ] . observe that for any @xmath175,$ ] the initial conditions for are taken as @xmath176 since has constant coefficients , the support of the asymptotic root - counting measure of its solution is given by the well - known result of beraha - cahane - weiss @xcite . namely , it coincides with the set of all @xmath177 such that the characteristic equation @xmath178 has two solutions with respect to @xmath179 which have the same modulus and , additionally , this modulus is the largest among all three solutions of ) . from considerations of @xcite one can easily derive that , for any fixed @xmath164,$ ] this support is the union of three intervals starting at the origin and ending at the branching points of which are given by the equation : @xmath180 these branching points are located as shown in figure [ figtau ] . observe that if for @xmath164 $ ] , @xmath181 denotes the branching point lying on the positive half - axis , then it attains its maximum at @xmath182 and this maximum equals @xmath46 . as we mentioned before , in the case under consideration all supports of lie on three fixed rays through the origin . therefore , the density of the asymptotic distribution of the sequence @xmath183 is obtained by averaging the densities of over @xmath164 $ ] , see @xcite and @xcite . therefore , the support of the asymptotic root distribution for @xmath184 is the union of three intervals of length @xmath46 . of the solution to the recurrence relation for @xmath185 , @xmath182 and @xmath186 resp . , title="fig : " ] of the solution to the recurrence relation for @xmath185 , @xmath182 and @xmath186 resp . , title="fig : " ] of the solution to the recurrence relation for @xmath185 , @xmath182 and @xmath186 resp . , title="fig : " ] we have to show that proposition [ conj:2 ] holds asymptotically for any sequence @xmath187 satisfying the condition @xmath33 . the principal minors of @xmath188 given by : @xmath189 satisfy the recurrence @xmath190 @xmath191 where @xmath104 runs from @xmath108 to @xmath109 with the boundary conditions @xmath192 to obtain a converging sequence of root - counting measures one has to consider @xmath193 @xmath194 which is the recurrence for the principal minors of @xmath195 given by : @xmath196 when @xmath197 the family of recurrence relations converges to the earlier family valid in case @xmath198 due to the fact that @xmath33 . therefore the measure obtained by averaging of the family of root - counting measures for recurrence relations with constant coefficient is exactly the same as in case @xmath198 , i.e. , it coincides with @xmath45 . additionally observe that by a general result of @xcite , the support of the asymptotic root - counting measure of the sequence @xmath199 can only be smaller than that of @xmath45 and their cauchy transforms coincide outside the support of @xmath45 . since the support of @xmath45 is the union of three straight intervals through the origin the resulting asymptotic root - counting measure for the sequence @xmath200 in case @xmath33 coincides with @xmath45 . since the support of the limiting measure @xmath45 consists of three segments through the origin it is possible in principle to write down integral formulas for the density and the cauchy transform of @xmath45 similar to those presented in @xcite , @xcite and @xcite . in particular , in the complement to the support of @xmath45 its cauchy transform is given by @xmath201 where @xmath202 is the unique solution of satisfying @xmath203 and @xmath204 is its partial derivative with respect to @xmath49 . however it seems difficult to find either a somewhat explicit expression for @xmath205 or a linear differential operator with polynomial coefficients annihilating @xmath205 which in a similar case the authors succeeded to obtain in @xcite . ( observe that such an operator always exists since @xmath205 is a nilsson - class function , see @xcite . ) fix @xmath3 and , similarly to the previous section , introduce a double indexed sequence @xmath206 of the characteristic polynomials of the principal minors of @xmath207 , see . for any fixed @xmath5 , we can easily obtain the corresponding recurrence relation of length @xmath23 : @xmath208 @xmath209 the same scaling by @xmath44 of the eigenvalues of @xmath207 ( which we conjecture results in compactly supported limiting root - counting measure @xmath51 ) leads to @xmath210 where @xmath211 . observe that is the recurrence relation for the sequence of characteristic polynomials of the principal minors of @xmath212 given by @xmath213 assuming that @xmath197 we obtain that tends to following relation with constant coefficients : @xmath214 whose characteristic equation is given by @xmath215 typically the union of the supports of the asymptotic root - counting measures for the polynomial sequences depending on @xmath164 $ ] is larger than that of @xmath51 , see fig . [ figa1 ] . in such case we can only conclude that the corresponding cauchy transforms coincide in the complement to the larger support . for @xmath216 and @xmath217 resp . , title="fig : " ] for @xmath216 and @xmath217 resp . , title="fig : " ] for @xmath216 and @xmath217 resp . , title="fig : " ] together with 3 branching points given by . ] in the real @xmath218-plane . ] however , in case @xmath219 ( which fits the situation covered by the main result of @xcite , ) these supports coincide and one can obtain the density of @xmath51 by averaging the densities of . [ apositive ] if @xmath220 then for @xmath164,$ ] each support of the family is a real interval . their union coincides with the ( real ) interval connecting two rightmost branching points defined by , see fig [ figa=3 ] . using symbolic manipulations , one can check that all three branching points of with respect to @xmath179 satisfy the equation : @xmath221 to check that for @xmath222 and any @xmath164,$ ] all three solutions of are real , we calculate the discriminant of with respect to @xmath49 . again using symbolic manipulations , we get that this discriminant is given by : @xmath223 for @xmath175,$ ] the graph of @xmath224 in the real @xmath218-plane is presented in fig . [ figatau ] . one can easily check that the maximal value of @xmath5 on this graph is obtained when @xmath182 and is equal to @xmath225 . at the same time , for @xmath226 one can easily check that all three roots of are real . for @xmath227 ( left ) , @xmath228 ( right ) and @xmath229 ( down ) in the @xmath49-plane . larger dots are the endpoints of the support given by . , title="fig : " ] for @xmath227 ( left ) , @xmath228 ( right ) and @xmath229 ( down ) in the @xmath49-plane . larger dots are the endpoints of the support given by . , title="fig : " ] for @xmath227 ( left ) , @xmath228 ( right ) and @xmath229 ( down ) in the @xmath49-plane . larger dots are the endpoints of the support given by . , title="fig : " ] similarly to the case @xmath230 for any @xmath231 , one can represent the cauchy transform of @xmath51 in the complement to its support ( which is an interval explicitly given in lemma [ apositive ] ) as @xmath201 where @xmath202 is the unique solution of satisfying the condition @xmath232 ( in an appropriate domain in @xmath57 such presentation for the cauchy transform of @xmath51 is valid for any complex @xmath5 . ) to obtain the support of @xmath51 we argue as follows . assume that we have a ( sub)sequence @xmath233 of the eigenvalues of @xmath234 ( one for each @xmath11 ) converging to some finite limit denoted by @xmath235 . denote by @xmath236 the corresponding ( sub)sequence of eigenpolynomials . ( these eigenpolynomials are classically referred to as stieltjes polynomials and in special cases their properties are well studied in the literature . ) each @xmath237 satisfies its own differential equation : @xmath238 * first assumption . * we assume that if the subsequence @xmath239 has a finite limit then the sequence @xmath240 of the root - counting measures of @xmath236 also weakly converges , after appropriate scaling of @xmath53 , to some limiting measure @xmath241 whose support consists of finitely many compact curves and points . this assumption implies that the sequence of cauchy transforms of scaled @xmath240 converges to the cauchy transform of the limiting measure @xmath241 . the appropriate scaling of @xmath53 can be easy guessed from the latter equation . namely , substituting @xmath242 and dividing the above equation by @xmath243 , we get the following relation : @xmath244 with respect to the new independent variable @xmath245 . observe that the scaled logarithmic derivative @xmath246 is the cauchy transform of the root - counting measure of @xmath247 with respect to the variable @xmath245 . * second assumption . * assuming that the sequence @xmath240 of the root - counting measures of @xmath248 converges to @xmath249 we additionally assume that the sequences of the root - counting measures of its first and second derivatives converge to the same measure @xmath250 ( apparently this assumption can be rigorously proved by using the same argulments as presented in @xcite . ) under the above assumptions , using , we get that the cauchy transform @xmath251 of @xmath241 satisfies the quadratic equation : @xmath252 almost everywhere in @xmath57 . up to a variable change @xmath253 and @xmath254 the latter equation coincides with . * third assumption . * so far we presented a physics argument showing that if a ( sub)sequence @xmath239 of the eigenvalues of @xmath212 converges to some limit @xmath255 then there exists a probability measure @xmath241 whose cauchy transform satisfies almost everywhere in the @xmath245-plane . our final assumption is that the converse to the latter statement is true as well , i.e. , for each @xmath235 with the above properties there exists an appropriate subsequence @xmath239 converging to @xmath235 . ( at the moment we do not have a mathematically rigorous proof of this statement but such fact in a very similar situation was settled in @xcite . ) the next result is a special case of proposition 9 and theorem 12 in @xcite . [ lm : horiz ] there exists a signed measure @xmath256 whose cauchy trasform satisfies almost everywhere if and only if the set of critical horizontal trajectories of the quadratic differential @xmath257 contains all its turning points , i.e. , all roots of @xmath258 . ( here by a critical horizontal trajectory of a quadratic differential we mean its horizontal trajectory which starts and ends at the turning points . ) we know that the support of @xmath256 should include all the branching points of and consists of the critical horizontal trajectories of . [ crval ] the set of the critical values of the polynomial @xmath259 , i.e. , the set of all @xmath235 for which @xmath260 has a double root with respect to @xmath245 is given by the equation : @xmath261 straight - forward calculations . [ endpoints ] the endpoints of the support of @xmath51 are contained among the three roots of equation when @xmath182 , see fig . this equation coincides with where @xmath235 is substituted by @xmath49 . for any positive integer @xmath11 and generic values of parameter @xmath81 the roots of @xmath28 with respect to @xmath18 ( i.e. , the quasi - exactly solvable spectrum ) are simple . for any given @xmath11 , and any sufficiently large positive @xmath5 , these roots are real and distinct . the set @xmath262 of branching points of @xmath28 , i.e. , the set of all values of @xmath5 for which two eigenvalues coalesce , has cardinality @xmath71 . when plotted these branching points form a regular pattern in the complex plane shown in figures [ bbigtriangle ] and [ triangle10 ] . in this subsection we present our ( mostly ) numerical results and conjectures about the monodromy of the roots of @xmath28 when @xmath5 runs along different closed paths in the complement to @xmath62 in the @xmath5-plane . we start with the following statement . [ prop : asymp ] for any given @xmath11 , if @xmath266 with @xmath267 fixed , then the roots of @xmath268 divided by @xmath44 will be asymptotically uniformly distributed on the straight segment @xmath269 $ ] , see fig . [ figslopes ] . in particular , if @xmath5 runs over the circle @xmath270 $ ] for any sufficiently large @xmath271 then the resulting monodromy of roots of @xmath272 ( which are all real ) is the complete reversing of their order , i.e. , the leftmost and the rightmost roots change places , the second from the left and the second from the right change places etc . observe that if @xmath273 with @xmath274 very large then polynomial @xmath275 , comp . , is close to @xmath276 where @xmath276 is the characteristic polynomial of the tridiagonal matrix @xmath277 to make the situation more transparent , let us consider the sequence of characteristic polynomials of @xmath278 and of @xmath279 in other words , we are comparing the roots of @xmath276 divided by @xmath280 with that of @xmath276 divided by @xmath280 . the characteristic polynomials of the respective principal minors of @xmath278 and of @xmath281 satisfy the recurrences : @xmath282 and @xmath283 where @xmath284 and both recurrences have the standard boundary conditions : @xmath285 . as before @xmath286 and @xmath287 . observe now that , for any fixed @xmath11 and any @xmath288 , one can choose @xmath289 so large that ( due to the presence of @xmath290 in the denominator of the third term in ) each equation in for @xmath291 deviated from the corresponding equation in so little that @xmath292 can be made arbitrary small coefficientwise . now one can easily check by induction that bivariate polynomial @xmath293 is quasihomogeneous with weight @xmath108 for variable @xmath18 and weight @xmath294 for variable @xmath5 . when @xmath5 is positive , then @xmath293 is a real - rooted polynomial . since multiplication of @xmath18 by @xmath295 and multiplication of @xmath5 by @xmath296 multiplies the whole @xmath293 by a constant , the roots of @xmath293 with respect to @xmath18 lie for any fixed @xmath5 on the line through the origin whose slope is half of the slope of @xmath5 . consider the roots of @xmath276 with respect to @xmath49 . observe that if @xmath297 , then the spectrum of is @xmath298 . the same argument as above gives that for any @xmath299 the roots of @xmath276 will be equally spaced on the interval @xmath300 $ ] . since choosing @xmath289 sufficiently large , we can achieve that all roots of @xmath301 lie arbitrarily close to that of @xmath302 , and the latter roots are equally distributed on the interval @xmath303,$ ] the result follows . to describe our conjecture on monodromy of the spectrum , let us introduce a system of standard paths connecting a base point which we choose as a sufficiently large positive number with every branching point , see fig . [ paths ] . based on our numerical experiments , we see that @xmath62 has a triangular shape @xmath304 with points regularly arranged into columns and rows in @xmath57 . there are @xmath11 columns ( enumerated from left to right ) where the @xmath305-th column consists of @xmath306 branching points with approximately the same real part and there are @xmath11 rows ( enumerated from bottom to top ) where the @xmath307-th row consists of points with approximately the same imaginary part . we denote the branching points @xmath308 where @xmath309 is the row number and @xmath310 is the column number . fixing a base point @xmath311 connect @xmath312 with every @xmath313 by a " vertical hook @xmath314 , i.e. , move from @xmath312 vertically to the imaginary part of @xmath313 , then move horizontally to the left till you almost hit @xmath313 , then circumgo @xmath313 counterclockwise along a small circle centered at @xmath313 and return back to @xmath312 along the same path . conjecturally , along such a path one will never hit any other branching points unless @xmath313 lies on the real axis . in other words , the imaginary parts of all branching points except for the real ones are all distinct . in case when @xmath313 is real one can slightly deform the suggested path ( which is a real interval ) in an arbitrary way to move it away from the real axis . the resulting monodromy will ( conjecturally ) be independent of any such small deformation , see below . [ conj : monod ] for any @xmath308 and any sufficiently large positive base point @xmath312 , the monodromy corresponding to the standard path @xmath314 is a simple transposition @xmath315 of the roots of @xmath316 ( which are all real ) ordered from left to right . this conjecture has been numerically checked by the second author for all @xmath317 . observe that since the system of standard paths gives a basis of the fundamental group @xmath318 , then knowing the monodromy for the standard paths , one can calculate the monodromy along any loop in @xmath319 based at @xmath312 . [ [ asymptotic - distribution - of - branching - points - of - sp_nalambda - and - yablonskii - vorobev - polynomials ] ] asymptotic distribution of branching points of @xmath28 and yablonskii - vorobev polynomials ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ there are two natural asymptotic regimes for the sequence @xmath79 when @xmath32 . in the first regime we just consider the sequence @xmath79 and study the stabilization of its portion in any fixed bounded domain of @xmath57 when @xmath32 . in the second regime we consider the sequence @xmath321 and study the limiting density of zeros in the triangular shape @xmath80 which is more and more densely filled by the roots of @xmath322 when @xmath32 . numerical experiments show that both limits exist . [ conj : scaling ] the sequences @xmath325 and @xmath326 when @xmath327 asymptotically fills in the curvilinear triangular shape @xmath80 , see fig [ bbigtriangle ] . the interior of @xmath80 is the set of all values @xmath3 for which the support of @xmath51 consists of three smooth segments = legs with a common point , see fig . [ figa ] ( left ) . the complement of @xmath80 consists of all values @xmath3 for which the support of @xmath51 is a single smooth segment , see fig . [ figa ] ( down ) . the boundary of @xmath80 consists of those @xmath3 for which the support of @xmath51 is a single curve with a singularity , i.e. one of the three legs disappears , see fig . [ figa ] ( right ) . s. beraha , j. kahane , n. j. weiss , _ limits of zeros of recursively defined families of polynomials _ , in `` studies in foundations and combinatorics '' , pp . 213232 , advances in mathematics supplementary studies vol . 1 , ed . rota , academic press , new york , 1978 .	motivated by the earlier results of @xcite and @xcite , we study theoretically and numerically the asymptotics and the monodromy of the quasi - exactly solvable part of the spectrum of the quasi - exactly solvable quartic introduced by c. m. bender and s. boettcher @xcite . in particular , we formulate a conjecture on the coincidence of the asymptotic shape of the configuration of the branching points of the latter quartic with the asymptotic shape of zeros of the yablonskii - vorobev polynomials recently described in @xcite@xcite and present its ( conjectural ) alternative description , see conjecture [ conj : scaling ] .
astronomical polarimetry requires frequent calibration operations to remove the instrumental polarization introduced by the various optical elements encountered by the light beam along its path . this contamination is usually determined by placing calibration optics early in the light path , which is used to feed light in a known state of polarization into the instrument . by measuring the polarization of the light that comes out of the system , it is possible to characterize it in terms of its jones matrix ( or , alternatively , the mueller matrix if one is using the stokes formalism ) . obviously , we can only characterize and calibrate the optical elements that the beam encounters _ after _ the polarization calibration optics . therefore , calibrating a telescope requires placing such optics at the telescope entrance , before the first reflection on the primary mirror ( m1 ) occurs . this approach@xcite is being successfully employed for the dunn solar telescope ( at the sacramento peak observatory , managed by the national solar observatory ) and the german vtt on the island of tenerife ( at the observatorio del teide of the instituto de astrof ' isica de canarias ) . in both cases , an array of linear polarizers and retarders are slided in the light path , on top of the entrance window , for calibration operations . separate calibrations are obtained for the telescope and the instrument , so that the former does not need to be done as frequently . the jones matrix of the complete system is then obtained as the product of the telescope and instrument matrices . unfortunately , entrance window polarizers are not practical for apertures larger than @xmath01 m diameter . in the past this has not been a major concern because : a)solar telescopes have apertures that do not exceed 1 m ; and b)large astronomical telescopes have not been used for polarization measurements . however , this scenario is starting to change . polarimetry is proving to be a very powerful tool to explore a broad range of astrophysical problems@xcite , resulting in a rapidly increasing interest to develop polarimeters for existing large telescopes . second , the development of the advanced technology solar telescope ( atst)@xcite demands a reliable method to calibrate a large telescope for high accuracy spectro - polarimetry . solar telescopes have been calibrated in the past by observing magnetic structures and making assumptions on the underlying physics . this poses important challenges , however , especially when pushing the envolope towards new observational domains . consider for example the atst , which is intended to do polarimetry at the @xmath010@xmath1 accuracy level . one of the common assumptions that is usually made in solar polarimetry is to consider that the continuum radiation is unpolarized . however , scattering processes can polarize the continuum and generate signals of the order of @xmath01% in the blue side of the visible spectrum . it has been stated @xcite that `` the direct observation of the polarisation of the continuous radiation still is a major outstanding observational challenge '' . considerations on the symmetry properties of stokes profiles are not appropriate either . gradients in the line - of - sight velocity or magnetic field introduce spurious asymmetries that can invalidate these assumptions . furthermore , some physical processes operating in the atoms are known to induce stokes asymmetries even in the absence of gradients . spectral lines forming in the incomplete paschen - back regime become asymmetric@xcite . moreover , the alignment - to - orientation conversion mechanism@xcite may also cause profile asymmetries . in summary , it is very important to have an absolute calibration that does not rely on preconceived ideas on the objects under study . this is specially true when the instrumentation is being used to explore new scientific realms . the present work has been motivated by the challenge of calibrating the 4-meter primary mirror of the atst to meet its very stringet polarization requirements . it might also be possible to use the calibration method proposed here in other existing large - aperture telescopes . however , the actual design of a practical implementation is beyond the scope of this paper . the main point of this work is to show that a calibration setup with inclined beam incidence can be used to measure the polarization properties of a large - aperture primary mirror . geometrical effects can be calculated and removed from the measured jones or mueller matrices resulting in a good approximation to such matrices in normal observing conditions . in this paper we shall consider two different configurations : the normal observing setup ( os ) and the calibration setup ( cs ) proposed here . fig [ fig : setups ] shows a schematic representation of the os and cs for an on - axis m1 mirror . our ultimate goal is to determine the jones matrix of m1 in the os ( @xmath2 ) . a direct measurement of @xmath2 would require polarization optics of the same diameter as the telescope aperture , which is not practical due to technical difficulties . the @xmath3 matrix , on the other hand , can be determined by mounting appropriate calibration optics ( and a mechanical control system ) at a height @xmath4 over m1 ( as shown in fig [ fig : setups ] , right ) . a small aperture on the dome is probably an ideal location for it . the cs requires some ( small ) amount of additional optics with respect to the os . at least two lenses are required : one at the entrance to open the beam and another at the detector for imaging . these additional elements should be designed with stringent polarization requirements . aberrations , chromatism and other image imperfections can be tolerated in these components , which will only be used for calibration . even if they introduce some residual polarization , this should be easily measurable in the laboratory and removed from the m1 jones matrix . the size of the calibration optics affects the accuracy of the calibration . in the limit where it fills the entire telescope aperture , the cs and the os are identical and no correction is needed . as the calibration aperture becomes smaller , the incidence angles increase resulting in larger corrections ( and errors ) . the simulations presented in this paper consider the pessimistic limit in which the calibration optics has a diameter approaching zero ( a pinhole aperture ) . for the calculations in this paper it is convenient to use cylindrical coordinates with the vertical axis along the propagation direction of the incident light beam . the radial coordinate @xmath5 is the distance from the center of the aperture and @xmath6 is the azimuth angle measured from an arbitrary reference . any given point on the surface of the m1 mirror at coordinates ( @xmath5 , @xmath6 ) is characterized by its complex refraction index @xmath7 . we shall consider here the behavior of a monochromatic plane wave . this will allow us to describe the instrumental polarization of m1 in terms of its jones matrix @xmath8 . appendix [ sec : appendix ] gives the mueller equivalent of the most important jones matrices derived in this work . locally , the behavior of m1 in the vicinity of ( @xmath5 , @xmath6 ) may be approximated by a reflection on a flat mirror of homogeneous refraction index @xmath9 . this process adopts a very simple form in the reference system of the plane of incidence ( the plane formed by the incoming ray and the surface normal , see fig [ fig : angles ] ) . let us denote the components in the plane of incidence with the subindex @xmath10 and those perpendicular to it with @xmath11 . in this frame , the jones matrix of the reflection , @xmath12 , is simply@xcite : @xmath13 assuming that the refraction index of air is 1 , @xmath14 and @xmath15 are given by : @xmath16 @xmath17 where @xmath18 is the angle of incidence . this matrix can be transformed to the global reference frame ( @xmath5 , @xmath6 ) : @xmath19 where @xmath20 is the usual rotation matrix . writing down @xmath8 explicitly : @xmath21 consider an axi - symmetric mirror illuminated by a collimated beam ( os ) . the angle of incidence @xmath18 is constant along concentric rings in the mirror ( @xmath22 ) . for example , a parabolic mirror of focal length @xmath23 is characterized by the condition : @xmath24 suppose that the complex refraction index is constant over the surface of the mirror ( perfect mirror ) . in this case , @xmath14 and @xmath15 are only functions of @xmath5 ( because @xmath22 ) . the dependences of @xmath8 in eq ( [ explicit ] ) are easily separable . the jones matrix of a thin ring of radius @xmath5 is simply : @xmath25 eq ( [ ring ] ) represents the jones matrix of a non - polarizing system ( identical reflectivity and retardance for both components of the electric field ) . this is a well - known property of axi - symmetric systems . notice , however , that the symmetry is broken if one observes away from the center of the field of view . imperfections in the mirror ( irregularities in the refraction index caused by coating degradation , dust , etc ) may also break the symmetry of the system and introduce instrumental polarization . let us now turn to the more general case of an imperfect mirror , defined as one with @xmath26 . if this is the case then @xmath14 and @xmath15 vary across the mirror and eq ( [ ring ] ) is no longer valid . we seek to determine a suitable calibration by means of the ( measurable ) @xmath3 matrix . the differences between @xmath2 and @xmath3 are due to the different incidence angle of the beam ( represented in fig [ fig : angles ] ) . we can expand @xmath14 in a power series of @xmath27 as : @xmath28 and similarly for @xmath15 . inserting this expansion into eq ( [ explicit ] ) we obtain : @xmath29 where @xmath30 and @xmath31 have been introduced for notational simplicity : @xmath32 ( and similarly for @xmath31 ) . in the equations above , @xmath27 , @xmath30 and @xmath31 are all functions of @xmath33 . the angle @xmath27 can be easily determined from geometrical considerations . however , @xmath30 and @xmath31 are affected by imperfections in the mirror that change over time . let us separate @xmath30 into two components : a nominal @xmath34 derived from eqs ( [ rp ] ) and ( [ dp ] ) with a theoretical refraction index @xmath35 ( e.g. , from manufacturer specifications ) , and an unknown @xmath36 due to coating degradation , dust accumulation , etc : @xmath37 ( and similarly for @xmath31 ) . inserting this into eq ( [ expandm ] ) and integrating over the entire mirror surface , we have : @xmath38 where @xmath39 is the radius of the m1 mirror . @xmath40 can be calculated numerically as the integral of @xmath41 : @xmath42 mirror imperfections are accounted for by the ( measured ) @xmath3 , whereas non - collimated incidence is accounted for by the ( calculated ) @xmath40 . @xmath43 is an unknown second - order term that couples mirror imperfections and non - collimated incidence . this term is small ( as shown below ) and may be neglected for our purposes here . the number of terms to retain in the taylor expansion of eq ( [ deltam ] ) depends on the particular telescope configuration and the accuracy required . typical examples are presented below in which @xmath40 can be neglected entirely ( on - axis mirror ) or needs to be calculated up to second order ( off - axis mirror , see [ sec : offaxis ] ) . in the reminder of this section i present the results of numerical simulations that provide some insight into the various terms that are involved in the calibration procedure . the parameters of the simulation are listed in table [ table : onaxis ] . they represent a 4-m on - axis telescope with a silver coating on the m1 mirror . the coating has been degraded so that the complex refraction index fluctuates over the mirror surface as @xmath44 $ ] ( @xmath45 is the average refraction index and @xmath46 is the amplitude of the fluctuation ) . this choice has been made to represent a pessimistic scenario that induces a considerable amount of instrumental polarization . the discretization of the simulation considers 100 points in @xmath5 and 200 in @xmath6 . the jones matrices @xmath2 and @xmath3 , obtained by applying eq ( [ explicit ] ) to each area element of the mirror , are : @xmath47 \ , , \ ] ] @xmath48 the `` irregularities '' introduced in the refraction index of the mirror break the symmetry and give rise to polarizing effects in @xmath2 , with off - diagonal terms of approximately 6% . fortunately , the calibration setup matrix @xmath3 is an excellent approximation to @xmath2 , with a maximum difference of @xmath010@xmath49 . it is then possible to calibrate the telescope almost down to the @xmath50 level without even having to correct for the inclined incidence ( i.e. , neglecting @xmath40 in eq [ [ expandm2 ] ] ) an off - axis mirror can be represented by a larger `` equivalent '' on - axis mirror with a variable refraction index , as depicted in fig [ fig : offaxis ] . the equations derived in [ sec : onaxis ] above are still valid in the @xmath33 reference frame . one simply needs to set the reflectivity to zero outside the shaded area of the figure ( @xmath51 ) . this can be accomplished , e.g. by setting the refraction index to 1 . this type of mirrors is slightly more complicated to calibrate . even a perfectly coated mirror will produce instrumental polarization . a simulation with the parameters listed in table [ table : offaxis ] yields the following matrices : @xmath52 \ , , \ ] ] @xmath53 @xmath3 is now significantly different from @xmath2 and we need to calculate the correction term @xmath40 ( eqs [ [ expandm2 ] ] and [ [ deltam ] ] ) . let us consider for the moment that the refraction index is perfectly known ( @xmath54 in eq [ [ splitd ] ] ) . calculating @xmath40 with eq ( [ deltam ] ) up to second order we have that : @xmath55 let us now turn to the more general case of an off - axis mirror with surface irregularities for which we have only an imperfect knowledge of the average refracion index . again , we use a refracion index with an angular dependence @xmath44 $ ] . furthermore , we do not know exactly the average refraction index of the mirror @xmath45 , but only an approximation @xmath35 . this approximate value will be used in the calculation of @xmath56 and @xmath40 ( eqs [ [ splitd ] ] and [ [ deltam ] ] ) . i carried out several experiments with different values of @xmath35 and @xmath57 to determine the sensitivity of the calibration to these parameters . some of the calculations include only the first - order dependence of @xmath40 on @xmath27 ( first term in the right - hand side of eq [ [ deltam ] ] ) . these are denoted by @xmath58 ( as opposed to @xmath40 , which considers the second - order term ) . table [ table : offaxis2 ] lists the results of these experiments . the first two rows give the highest polarizing term in @xmath2 for each simulation . the third to fifth rows show how the calibration error decreases with successive levels of approximation . by comparing the second and fifth rows , we can see that the calibration is able to reduce the instrumental polarization by an amount between one and two orders of magnitude . this paper introduces a new concept to calibrate telescopes for astronomical polarimetry . the proposed method is particularly useful for modern large - aperture telescopes , for which it is probably the only practical procedure ( at least for purely instrumental calibration ) . an accurate absolute calibration will be crucial for the new weak - signal science that the atst will open . existing night - time telescopes may also take advantage of this calibration procedure . the jones ( and mueller ) matrix of an on - axis mirror is almost unaffected by the non - collimated incidence of the beam in the cs . for the particular configuration considered in [ sec : onaxis ] , the jones matrix obtained from the calibration is good to almost @xmath50 with no need for any additional correction ( see eq [ [ onmcs ] ] ) . this is in spite of the relatively large mirror imperfections in the simulation , which generate polarizing terms of the order of 6% . an off - axis mirror suffers more instrumental polarization due to the asymmetric configuration . a mirror that produces instrumental polarization of a few percent can be calibrated to reach the @xmath59 level ( see table [ table : offaxis2 ] ) . in addition to measuring the jones matrix of the calibration setup ( @xmath3 ) , it is also necessary to calculate the correction term @xmath40 . this calculation is straightforward , though , and @xmath40 does not need to be recalculated unless the mirror is recoated or it degrades to a point where its _ average _ refraction index changes significantly . note that the @xmath59 calibration accuracy includes some uncertainty on the average refraction index of the mirror . i have used in this paper the jones matrix formalism , which deals directly with the components of the electric field of the light wave . sometimes , however , the mueller formalism is more adequate , especially when dealing with partially polarized or non - monochromatic light . many researchers are more familiar with the mueller matrices and the stokes parameters . for these reasons it is probably useful to provide the mueller equivalent@xcite of the jones matrices derived in this work . eq ( [ onmos ] ) becomes : @xmath60 eq ( [ onmcs ] ) becomes : @xmath61 eq ( [ offmos ] ) becomes : @xmath62 eq ( [ offmcs ] ) becomes : @xmath63 eq ( [ offmcsdm ] ) becomes : @xmath64 this work utilizes data from the advanced technology solar telescope ( atst ) project , managed by the national solar observatory , which is operated by aura , inc . under a cooperative agreement with the national science foundation . a. gandorfer , `` the second solar spectrum in the ultraviolet , '' in _ solar polarization workshop 3 _ , vol 307 of astronomical society of the pacific conference series , j. trujillo bueno and j. snchez almeida , eds . , asp , san francisco , california ) , p. 399 ( 2003 ) . h. socas - navarro , j. trujillo bueno , and e. landi deglinnocenti , `` `` signatures of incomplete paschen - back splitting in the polarization profiles of the hei 10830 multiplet '' , '' * _ in press _ * ( 2004 ) . e. landi deglinnocenti , in _ astrophysical spectropolarimetry _ , j. trujillo bueno , f. moreno - insertis , and f. s ' anchez , eds . , xii canary islands winter school of astrophysics , p. 1 ( cambridge university press , cambridge , uk , 2002 ) . max(\|m^os - m^cs\| ) & 1.2510 ^ -2 & 1.2510 ^ -2 & 1.2610 ^ -2 & 1.2610 ^ -2 + max [ \|m^os- + ( m^cs + m^*)\| ] & 4.0410 ^ -3 & 3.89 10 ^ -3 & 3.7110 ^ -3 & 2.5910 ^ -3 + max [ \|m^os- + ( m^cs + m)\| ] & 2.8410 ^ -4 & 4.1810 ^ -4 & 7.1110 ^ -4 & 1.8010 ^ -3 +	this paper describes a concept for the high - accuracy absolute calibration of the instrumental polarization introduced by the primary mirror of a large - aperture telescope . this procedure requires a small aperture with polarization calibration optics ( e.g. , mounted on the dome ) followed by a lens that opens the beam to illuminate the entire surface of the mirror . the jones matrix corresponding to this calibration setup ( with a diverging incident beam ) is related to that of the normal observing setup ( with a collimated incident beam ) by an approximate correction term . numerical models of parabolic on - axis and off - axis mirrors with surface imperfections are used to explore its accuracy .
we report here on an important step towards the ultimate goal of constructing numerical relativity codes that calculate accurately in 3d the gravitational radiation at future null infinity . by `` accurately '' we mean ( at least ) second - order convergent to the true analytic solution of a well posed initial value problem . thus our goal is to provide an accurate and unambiguous computational map from initial data to gravitational waveforms at infinity . of course , uncertainties will always exist in the appropriate initial data for any realistic astrophysical system ( e.g. in a binary neutron star system , the data for the metric components would not be uniquely determined by observations ) . but such a computational map enables focusing on the underlying physics in a rigorous way . most relativity codes are second - order convergent , but because of boundary problems the convergence may not be to the true analytic solution of the intended physical problem . in order to explain this point , and to give the idea behind our method , we first briefly review some aspects of numerical relativity . the predominant work in numerical relativity is for the cauchy `` 3 + 1 '' problem , in which spacetime is foliated into a sequence of spacelike hypersurfaces . these hypersurfaces are necessarily of finite size so , in the usual case where space is infinite , an outer boundary with an artificial boundary condition must be introduced . this is the first source of error because of artificial effects such as the reflection of outgoing waves by the boundary . next , the gravitational radiation is estimated from its form inside the boundary by using perturbative methods , which ignore the nonlinear aspects of general relativity in the region outside the boundary . for these reasons the numerical estimate of gravitational radiation is not , in general , convergent to the true analytic value at future null infinity . the radiation properties of the robinson - trautman metric will be used to illustrate this effect . an alternative approach in numerical relativity uses the characteristic formalism , in which spacetime is foliated into a sequence of null cones emanating from a central geodesic . this approach has the advantage that the einstein equations can be compactified @xcite so that future null infinity is rigorously represented on a finite grid , and there is no artificial outer boundary condition . however , it suffers from the disadvantage that the coordinates are based on light rays , which can be focussed by a strong field to form caustics which complicate a numerical computation @xcite . also , to date , the characteristic initial value problem has only been implemented numerically for special symmetries @xcite . our ultimate goal is a 3d cauchy - characteristic matching ( ccm ) code , which uses the positive features of the two methods while avoiding the problems . more precisely , the interior of a timelike worldtube @xmath0 is evolved by a cauchy method , and the exterior to future null infinity is evolved using a characteristic algorithm ; boundary conditions at @xmath0 are replaced by a two - way flow of information across @xmath0 . in relativity , under the assumption of axisymmetry without rotation , there has been a feasibility study of ccm @xcite ; see also @xcite . ccm has been successfully implemented for nonlinear wave equations and demonstrated to be second - order convergent to the true analytic solution ( which is not true in a pure cauchy formulation with sommerfeld outer boundary condition ) @xcite . while ccm has aesthetic advantages , it is important to ask whether it is an efficient approach . the question can be posed as follows . for a given target error @xmath1 , what is the amount of computation required for ccm compared to that required for a pure cauchy calculation ? it will be shown that the ratio tends to @xmath2 as @xmath3 , so that in the limit of high accuracy the effort is definitely worthwhile @xcite . our first step towards ccm is cauchy - characteristic extraction ( cce ) and we will present a partial implementation of cce in this paper . the idea of cce is to run a pure cauchy evolution with an approximate outer boundary condition . a worldtube @xmath0 is defined in the interior of the cauchy domain , and the appropriate characteristic data is calculated on @xmath0 ; then characteristic algorithms are used to propagate this gravitational field to future null infinity @xcite . cce is simpler than ccm to implement numerically , because in cce the data flow is one - way ( cauchy to characteristic ) whereas in ccm the data flows in both directions . note that the advantage of computational efficiency applies only to ccm and not to cce . however , we will show that the advantage of second - order convergence to the true analytic solution does apply , under certain circumstances , to cce . the work in this paper is part of the binary black hole grand challenge , which is concerned with the gravitational radiation resulting from the in - spiral and coalescence of two arbitrary black holes . however , the methods described here are not limited to black hole coalescence and could be applied to gravitational radiation from any isolated system , either with or without matter . in sec . [ sec:2 ] , we present a formalism for 3d characteristic numerical relativity in which the coordinates are based on null cones that emanate from a timelike worldtube @xmath0 ( recall that existing codes are in 2d with null cones emanating from a timelike geodesic ) @xcite . the characteristic einstein equations are written as a sum of two parts : quasispherical ( in a sense defined below ) plus nonlinear . the discretization and compactification of the einstein equations , with the nonlinear part ignored , is discussed in sec . [ sec:3 ] . a computer code has been written and in sec . [ sec:4 ] this code is tested on linearized solutions of the einstein equations , and extraction is tested on the nonlinear robinson - trautman solutions . the robinson - trautman solutions are also used to investigate the error of perturbative methods in estimating the gravitational radiation at null infinity . [ sec:5 ] uses the formalism developed in sec . [ sec:2 ] to estimate the errors associated with the finite boundary in a pure cauchy computation . this leads to the result concerning computational efficiency of ccm stated above . in the conclusion we discuss the further steps needed for a full implementation of cce , and also of ccm , and investigate under what circumstances cce can provide second - order convergence to the true analytic solution at future null infinity . we finish with appendices on the null cone version of gauge freedom and linear solutions of the einstein equations , and on a stability analysis of our algorithm . this is the first step towards a 3d characteristic evolution algorithm for the fully nonlinear vacuum einstein equations . here we treat the quasispherical case , where effects which are nonlinear in the asymmetry can be ignored . thus the schwarzschild metric is treated exactly in this formalism . however , rather than developing an algorithm for the linearized equations on a given schwarzschild background , we will approach this problem in a mathematically different way . we adopt a metric based approach in which each component of einstein s equation has ( i ) some quasispherical terms which survive in the case of spherical symmetry and ( ii ) other terms which are quadratic in the asymmetry , i.e terms of @xmath4 where @xmath5 measures deviation from spherical symmetry . we will treat the quasispherical terms to full nonlinear accuracy while discarding the quadratically asymmetric terms . for example , if @xmath6 were a scalar function we would make the approximation @xmath7 although this breakup is not unique , once made it serves two useful purposes . first , the resulting field equations are physically equivalent to the linearized einstein equations in the quasispherical regime . ( in the exterior vacuum region , the spherical background must of course be geometrically schwarzschild but the quasispherical formalism maintains arbitrary gauge freedom in matching to an interior solution ) . second , the resulting quasispherical evolution algorithm supplies a building block which can be readily expanded into a fully nonlinear algorithm by simply inserting the quadratically asymmetric terms in the full einstein equations . we use coordinates based upon a family of outgoing null hypersurfaces . we let @xmath8 label these hypersurfaces , @xmath9 ( @xmath10 ) , be labels for the null rays and @xmath11 be a surface area distance . in the resulting @xmath12 coordinates , the metric takes the bondi - sachs form @xcite @xmath13 where @xmath14 and @xmath15 , with @xmath16 a unit sphere metric . later , for purposes of including null infinity as a finite grid point , we introduce a compactified radial coordinate . a schwarzschild geometry is given by the choice @xmath17 , @xmath18 , @xmath19 and @xmath20 . to describe a linear perturbation , we would set @xmath21 and would retain only terms in @xmath22 which were of leading order in the linearization parameter @xmath5 . here we take a different approach . we express @xmath23 in terms of a complex dyad @xmath24 ( satisfying @xmath25 , @xmath26 , @xmath27 , with @xmath28 ) . then the dyad component @xmath29 is related to the linearized metric by @xmath30 . in linearized theory , @xmath31 would be a first order quantity . the 2-metric is uniquely determined by @xmath31 , since the determinant condition implies that the remaining dyad component @xmath32 satisfies @xmath33 . refer to @xcite for further details , especially how to discretize the covariant derivatives and curvature scalar of a topologically spherical manifold using the @xmath34 calculus . because the 2-metric also specifies the null data for the characteristic initial value problem , this role can be transferred to @xmath31 . terms in einstein equations that depend upon @xmath31 to higher than linear order are quadratically asymmetric . we do not explicitly introduce @xmath5 as a linearization parameter but introduce it where convenient to indicate orders of approximation . the einstein equations @xmath35 decompose into hypersurface equations , evolution equations and conservation laws . in writing the field equations , we follow the formalism given in @xcite . we find : @xmath36 where @xmath37 is the covariant derivative and @xmath38 the curvature scalar of the 2-metric @xmath39 . the quasispherical version of ( [ eq : beta ] ) follows immediately from rewriting it as @xmath40 , where @xmath41 is quadratically asymmetric . this defines the quasispherical equation @xmath42 thus in this approximation , @xmath43 . for a family of outgoing null cones which emanate from a nonsingular geodesic worldline , we could choose coordinate conditions so that @xmath44 . similarly , in minkowski space , we could set @xmath44 for null hypersurfaces which emanate from a non - accelerating spherical worldtube of constant radius . in a schwarzschild spacetime , due to red shift effects , @xmath45 need not vanish even on a spherically symmetric worldtube . thus @xmath45 represents some physical information as well as asymmetric gauge freedom in the choice of coordinates and choice of world tube . we wish to apply the same procedure to equations ( [ eq : u ] ) and ( [ eq : v ] ) . in doing so , it is useful to introduce the @xmath46 tensor field @xmath47 which represents the difference between the connection @xmath37 and the unit sphere connection @xmath48 , e.g. @xmath49 . in solving for @xmath50 , we use the intermediate variable @xmath51 then ( [ eq : u ] ) reduces to the first order radial equations @xmath52 we deal with these equations in terms of the spin - weighted fields @xmath53 and @xmath54 . to obtain quasispherical versions of these equations , we rewrite ( [ eq : qa ] ) and ( [ eq : ua ] ) as @xmath55 where @xmath56 , \\ n_u & = & r^{-2}e^{2\beta}q_{a } \left(h^{ab}-q^{ab}\right ) q_b.\end{aligned}\ ] ] the quasispherical versions obtained by setting @xmath57 in ( [ eq : wqa ] ) and @xmath58 in ( [ eq : wua ] ) then take the form @xmath59 in terms of the spin - weighted differential operator @xmath34 . since @xmath60 and @xmath61 are asymmetric of @xmath46 , we use the gauge freedom to ensure that @xmath62 and @xmath63 are @xmath46 . since @xmath64 in minkowski space , we set @xmath65 in terms of a quasispherical variable @xmath66 . then ( [ eq : v ] ) becomes @xmath67 where @xmath68 - \frac{1}{4}r^4 e^{-2\beta } h_{ab } u^a_{,r } u^b_{,r}. \label{eq : nw}\ ] ] we set @xmath69 in ( [ eq : ww ] ) to obtain the quasispherical field equation for @xmath70 . next , by the same procedure , the evolution equations take the form @xmath71 where @xmath72 . \label{eq : nj}\end{aligned}\ ] ] the quasispherical evolution equation follows from ( [ eq : wev ] ) by setting @xmath73 . the remaining independent equations are the conservation conditions . for a worldtube given by @xmath74 , these are given in terms of the einstein tensor by @xmath75 where @xmath76 is any vector field tangent to the worldtube . this expresses conservation of @xmath77-momentum flowing across the worldtube @xcite . these equations simplify when the bondi coordinates are adapted to the worldtube so that the angular coordinates @xmath9 are constant along the @xmath78 streamlines . then @xmath79 on the worldtube and an independent set of conservation equations is given ( in the quasispherical approximation ) in terms of the ricci tensor by @xmath80 in the context of an extraction problem it is assumed that the interior solution satisfies the einstein equations , and therefore that the conservation conditions are automatically satisfied on the extraction worldtube . the above equations define a quasispherical truncation of the vacuum einstein equations . because these quasispherical equations retain some terms which are nonlinear in the asymmetry , their solutions are not necessarily linearized solutions in a schwarzschild background . however , in the perturbative limit off schwarzschild , the linearized solutions to these truncated equations agree with the linearized solutions to the full einstein equations . in this section we describe a numerical implementation , based on second - order accurate finite differences , of the equations presented in sec . [ sec:2 ] . we introduce a compactified radial coordinate @xmath81 , ( with @xmath82 being the extraction radius ) , labeling null rays by the real and imaginary parts of a stereographic coordinate @xmath83 on the sphere , i.e. @xmath84 . the radial coordinate is discretized as @xmath85 for @xmath86 and @xmath87 . here @xmath88 defines a world tube of constant surface area coordinate . the point @xmath89 lies at null infinity . the stereographic grid points are given by @xmath90 and @xmath91 for @xmath92 and @xmath93 . the fields @xmath31 , @xmath94 , @xmath62 and @xmath66 are represented by their values on this rectangular grid , e.g. @xmath95 . however , for stability ( see appendix [ app : stab ] ) , the field @xmath63 is represented by values at the points @xmath96 on a radially staggered grid ( accordingly @xmath97 ) . for the extraction problem , it is assumed that the values of the fields and the radial derivative of @xmath63 are known at the boundary . in the following discussion , it is useful to note that asymptotic flatness implies that the fields @xmath98 , @xmath99 , @xmath100 and @xmath101 are smooth at @xmath102 , future null infinity @xmath103 . in terms of the compactified radial variable @xmath104 , the quasispherical field equation for @xmath62 reduces to @xmath105 we write all derivatives in centered , second order accurate form and replace the value @xmath106 by its average @xmath107 . the resulting algorithm determines @xmath108 in terms of values of @xmath31 and @xmath94 at the points @xmath109 , @xmath110 and @xmath111 @xmath112 ( here and in what follows , we make explicit only the discretization on the radial direction @xmath104 , and we suppress the angular indices @xmath113 ) . since eq . ( [ eq : qdisc ] ) is a 3-point formula , it can not be applied at the second point , however , a suitable formula for @xmath114 is given by @xmath115 where the value of @xmath116 is trivially obtained from the knowledge of @xmath61 at the boundary , and @xmath117 , @xmath118 . after a radial march , the local truncation error compounds to an @xmath119 global error in @xmath62 at @xmath103 . in terms of the compactified radial variable @xmath104 , the quasispherical field equation for @xmath63 reduces to @xmath120 we again rewrite all derivatives in centered , second order form . because of the staggered placement of @xmath63 , the resulting discretization is @xmath121 the value of @xmath63 at the first point is evaluated from the expansion @xmath122 at the boundary . this leads to an algorithm for determining @xmath63 at the point @xmath123 in terms of values of @xmath62 at the points @xmath124 lying on the same angular ray . after completing a radial march , local truncation error compounds to an @xmath119 global error in @xmath63 at @xmath125 . the quasispherical field equation for @xmath66 ( [ eq : ww ] ) , reexpressed in terms of @xmath104 and @xmath126 , is @xmath127 following the same procedure as in eq . ( [ eq : qdisc ] ) we obtain @xmath128 we obtain a startup version of the above with the substitutions @xmath129 , @xmath130 , noting that at the boundary @xmath131 is given . the above algorithm has a local error @xmath132 in each zone . in carrying out the radial march , this leads to @xmath133 error at any given physical point in the uncompactified manifold . however , numerical analysis indicates an @xmath134 error at @xmath103 . in discretizing the evolution equation , we follow an approach that has proven successful in the axisymmetric case @xcite and recast it in terms of the 2-dimensional wave operator @xmath135 \label{eq:2box}\ ] ] corresponding to the line element @xmath136 , \label{eq:2metric}\ ] ] where @xmath137 is the normal to the outgoing null cones and @xmath138 is a null vector normal inwards to the spheres of constant @xmath11 . because the domain of dependence of @xmath139 contains the domain of dependence induced in the @xmath140 submanifold by the full space - time metric ( [ eq : bmet ] ) , this approach does not lead to convergence problems . the quasispherical evolution equation ( [ eq : wev ] ) then reduces to @xmath141 where @xmath142 because all 2-dimensional wave operators are conformally flat , with conformal weight @xmath143 , we can apply to ( [ eq : wev1 ] ) a flat - space identity relating the values of @xmath144 at the corners @xmath145 , @xmath62 , @xmath82 and @xmath146 of a null parallelogram @xmath147 with sides formed by incoming and outgoing radial characteristics . in terms of @xmath144 , this relation leads to an integral form of the evolution equation , @xmath148 the corners of the null parallelogram can not be chosen to lie exactly on the grid because the velocity of light in terms of the @xmath104 coordinate is not constant . numerical analysis and experimentation has shown @xcite that a stable algorithm results by placing this parallelogram so that the sides formed by incoming rays intersect adjacent @xmath8-hypersurfaces at equal but opposite @xmath104-displacement from the neighboring grid points . the elementary computational cell consists of the lattice points @xmath149 and @xmath150 on the `` old '' hypersurface and the points @xmath151 , @xmath152 and @xmath153 . the values of @xmath144 at the vertices of the parallelogram are approximated to second order accuracy by linear interpolations between nearest neighbor grid points on the same outgoing characteristic . then , by approximating the integrand by its value at the center @xmath154 of the parallelogram , we have @xmath155 as a result , the discretized version of ( [ eq : wev1 ] ) is given by @xmath156 where @xmath157 is a linear function of the @xmath158 s and angular indexes have been suppressed . consequently , it is possible to move through the interior of the grid computing @xmath159 by an explicit radial march using the fact that the value of @xmath144 on the world tube is known . the above scheme is sufficient for second order accurate evolution in the interior of the radial domain . however , for startup purposes , special care must be taken to handle the second radial point . in determining @xmath160 the strategy ( [ eq : wev2 ] ) is easily modified so that just two radial points are needed on the @xmath161 level ; the parallelogram is placed so that @xmath145 and @xmath62 lie precisely on @xmath162 and @xmath163 respectively . note that the calculation of @xmath164 poses no problems , since the values of @xmath66 , @xmath63 , and @xmath61 are known on the worldtube and the value of @xmath70 on the worldtube can be calculated by ( [ eq : ww ] ) . in order to apply this scheme globally we must also take into account technical problems concerning the order of accuracy for points near @xmath103 . for this purpose , it is convenient to renormalize ( [ eq : wev3 ] ) by introducing the intermediate variable @xmath165 . this new variable has the desired feature of finite behavior at @xmath103 . with this substitution the evolution equation becomes @xmath166 where all the terms have finite asymptotic value . some of the fundamental issues underlying stability of the evolution algorithm are discussed in appendix [ app : stab ] . we have carried out numerical experiments which confirm that the code is stable , subject to the cfl condition , in the perturbation regime where caustics and horizons do not form . the first set of tests consist of evolving short wavelength initial null data , with all world tube data set to zero . in this case , the world tube effectively acts as a mirror to ingoing gravitational waves . the tests were run until all waves were reflected and radiated away to @xmath103 . in particular , data with @xmath167 was run from from @xmath168 to @xmath169 , corresponding to approximately @xmath170 timesteps , at which time it was checked that the amplitude was decaying . in the second set of tests , we included short wavelength data with amplitude @xmath171 for the boundary values of @xmath94 , @xmath31 , @xmath63 , @xmath62 and @xmath66 on the world tube ( with compact support in time ) as well for the initial data for @xmath31 ( with compact support on the initial null hypersurface ) . again the code was run for approximately @xmath172 timesteps ( from @xmath168 to @xmath173 ) , at which time all fields were decaying exponentially . this test reveals a favorably robust stability of the worldtube initial value problem , since in this case the world tube conservation conditions which guarantee that the exterior evolution be a vacuum einstein solution were not imposed upon the worldtube data . we now present code tests for the accuracy of numerical solutions and their waveforms at infinity . the tests are based upon linearized solutions on a minkowski background and linearized robinson - trautman solutions . these solutions provide testbeds for code calibration as well as consistent worldtube boundary values for an external vacuum solution . in addition , we use numerical solutions of the nonlinear robinson - trautman equation to study the waveform errors introduced by the quasispherical approximation . appendices [ app : gauge ] and [ app : lin ] describe how to generate 3-dimensional linearized solutions on a minkowski background in null cone coordinates and their gauge freedom . to calibrate the accuracy of the code , we choose a solution of ( [ eq : wave ] ) and ( [ eq : ecr ] ) which represents an outgoing wave with angular momentum @xmath174 of the form @xmath175 where @xmath176 is the @xmath177-translation operator . the resulting solution is well behaved above the singular light cone @xmath168 . convergence was checked in the linearized regime by choosing initial data of very small amplitude @xmath178 . we used the linearized solution ( [ eq : exacsol ] ) to give data at @xmath179 , with the inner boundary at @xmath180 , and we compared the numerically evolved solution at @xmath181 . the computation was performed on grids of size @xmath182 equal @xmath183 , @xmath184 , @xmath185 and @xmath186 , while keeping @xmath187 . convergence to second order was verified in the @xmath188 , @xmath189 and @xmath190 norms . the robinson - trautman space - times @xcite contain a distorted black hole emitting purely outgoing radiation . the metric can be put in the bondi form @xmath191 where @xmath192 $ ] , @xmath193 is the angular momentum operator and @xmath194 satisfies the nonlinear equation @xmath195 the schwarzschild solution ( for a unit mass black hole ) is obtained when @xmath196 . more generally , smooth initial data @xmath197 evolves smoothly to form a schwarzschild horizon . the linearized solutions to the robinson - trautman equation ( [ eq : rteq ] ) are obtained by setting @xmath198 and dropping nonlinear terms in @xmath6 : @xmath199 for a spherical harmonic perturbation @xmath200 this leads to the exponential decay @xmath201 . these linearized solutions provide analytic worldtube data for our evolution code , along with the initial null data @xmath202 . we have used this as a check of code accuracy in the perturbative regime off schwarzschild . with this data , the code should evolve @xmath31 to be globally zero to second order in grid size . of particular importance for the extraction of waveforms , this should hold for the value of @xmath31 at @xmath103 . we have carried out such a test with a small extraction radius ( @xmath203 ) and a linearized solution of the form : @xmath204\ ] ] with @xmath205 . the error norm @xmath206 was determined by integration over a sphere at @xmath207 with solid angle element @xmath208 , and with an integration time of @xmath209 . the convergence rate to the true value was found to be @xmath210 . we have also obtained second - order accurate numerical solutions to the nonlinear robinson - trautman equation ( [ eq : rt ] ) . see ref . @xcite for numerical details . this allows us to check the discrepancy between exact waveforms and waveforms obtained by regarding the whole spacetime in the quasispherical perturbative approximation . we have based this comparison on initial data in modes @xmath211.\ ] ] in order to supply some physical perspective , the nonlinearity of the initial data is best measured in terms of @xmath212 , where @xmath213 is the initial mass of the system and @xmath214 is the mass of the corresponding schwarzschild background . ( here , @xmath215 ) . we also calculate the percentage of the initial mass which is radiated away during the entire course of our simulations . the bondi news function determines the mass loss and it is also an appropriate physical quantity to invariantly describe radiative waveforms . in the coordinates adopted here , the news function is given by @xcite @xmath216 where the bondi time @xmath217 measured by observers at @xmath125 is related to @xmath8 by @xmath218 . for various initial modes , we have calculated the news function for the numerical solution of the nonlinear r - t equation ( @xmath219 ) and compared it with the news function of the linearized solution ( @xmath220 ) . as expected , for small values of @xmath221 they agree up to second order in @xmath221 . figure [ fig : kuabi_2 ] graphs the time dependence of @xmath222 ( for a representative angle ) for a system initially in a @xmath223 , @xmath224 mode , which is the dominant gravitational radiation mode for a spiraling binary system . the figure illustrates that @xmath225 scales with @xmath226 . however , for larger @xmath221 , corresponding to a total radiative mass loss greater than @xmath227 , this is no longer the case and a noticeable discrepancy arises . for instance , as illustrated in figure [ fig : kuama_2 ] , the difference between quadratically rescaling @xmath225 and its actual value is about @xmath228 for a mass loss of @xmath229 . hence , this indicates that not only the first order perturbation treatment but also the second order treatment is grossly inaccurate in this regime . serious discrepancies arise between @xmath219 and @xmath220 for ranges in which the mass loss is not extreme . in fact , @xmath219 reveals an oscillatory behavior qualitatively quite different from the pure decaying mode of the perturbative solution , which has serious implications for the tidal acceleration which the radiation would produce in a distant gravitational wave antenna . as measured by the radiative component of the weyl tensor @xmath230 , the tidal acceleration is given by the bondi - time derivative of the news function . in contrast to the monotonic decay of the perturbative solution , the actual behavior of @xmath230 exhibits damped oscillations . for a @xmath231 initial mode , figure [ fig : radia ] shows the drastic difference between the numerically obtained @xmath232 and the corresponding @xmath233 calculated with the perturbative solution . similar nonlinear oscillations arise from other choices of initial data . some partial explanation of this behavior might be possible using second order perturbation theory for the robinson - trautman equation @xcite but the full behavior would require perturbation expansions far beyond practicality . this section is concerned with the computational efficiency of a numerical calculation of gravitational radiation from an isolated system , such as binary black hole . by `` computational efficiency '' we mean the amount of computation @xmath234 ( i.e. the number of floating point operations ) for a given target error @xmath1 . we will show that the computational efficiency of the ccm algorithm is never significantly worse than that of a pure cauchy algorithm ; and that for high accuracy the ccm algorithm is always much more efficient . in ccm a `` 3 + 1 '' interior cauchy evolution is matched to an exterior characteristic evolution at a worldtube @xmath235 . a key feature is that the characteristic evolution can be rigorously compactified , so that the whole space - time to future null infinity may be represented on a finite grid . from a numerical point of view this means that the only error made in a calculation of the gravitational radiation at infinity is due to the finite discretization @xmath236 ; for second - order algorithms this is @xmath119 . the value of the matching radius @xmath237 is important and it will turn out that , for efficiency , it should be as small as possible . however , caustics may form if @xmath237 is too small . the smallest value of @xmath237 that avoids caustics is determined by the physics of the problem , and is _ not _ affected by either the discretization @xmath236 or the numerical method . on the other hand , the standard approach is to make an estimate of the gravitational radiation solely from the data calculated in a pure cauchy evolution . the simplest method would be to use the raw data , but that approach is too crude because it mixes gauge effects with the physics . thus a substantial amount of work has gone into perturbative methods that factor out the gauge effects using multipole expansions and estimate the gravitational field at infinity from its behavior within the domain of the cauchy computation @xcite . we will call this method waveform extraction , or we . while we is a substantial improvement on the crude approach , it ignores the nonlinear terms in the einstein equations . the resulting error will be estimated below . both cce and we are _ extraction _ methods . that is , they use cauchy data on a worldtube @xmath0 to estimate gravitational waveforms at infinity , and they have no back effect on the cauchy evolution . in both methods there is an error ( which is difficult to estimate ) due to the artificial cauchy outer boundary condition . the difference between cce and we is in the treatment of the nonlinear terms between @xmath0 and future null infinity and in the truncation of the perturbative multipole expansion at some low order . we ignores the nonlinear terms , and this is an inherent limitation of a perturbative method . even if it is possible to extend we beyond linear order , there would necessarily be a cut - off at some finite order . the quasispherical implementation of cce incorporates all multipole contributions but also ignores the nonlinear terms . however , it is in principle straightforward to incorporate these terms into the code . a full implementation of cce would do so , and the nonlinear terms would be treated without error . we assume that a pure cauchy evolution proceeds in a spatial domain of radius @xmath238 , and the extraction is computed on a worldtube @xmath0 of radius @xmath82 , with @xmath239 . the evolution equation ( [ eq : wev ] ) may be written : @xmath240 with the nonlinear term @xmath241 given by ( [ eq : nj ] ) ( actually , @xmath241 also implicitly contains contributions from @xmath242 and @xmath243 , and from the quasispherical approximation of terms in ( [ eq : wev ] ) , but these effects are all of the same order as @xmath241 ) . the order of magnitude of various terms can be expressed in terms of a function @xmath244 ( whose time - derivative is the news function ) ; note that @xmath245 is not a small quantity . the expressions are : @xmath246 these estimates are obtained by the radial integration of the field equations in sec . [ sec:2.2 ] , assuming that the background geometry is minkowskian and that the bondi gauge conditions are satisfied . should this not be the case then constants of order unity would be added to @xmath62 , @xmath63 and @xmath66 , and the effect of this would be to amend ( [ eq : wev ] ) by adding terms to the quasispherical part so that it represents wave propagation on a ( fixed ) non - flat background . however , the order of magnitude of terms in the nonlinear part would not be affected . thus there is no loss of generality , and a significant gain in simplicity and transparency , in performing the error analysis on a minkowskian background . it is straightforward to confirm that the nonlinear correction to ( [ eq : wev ] ) involves terms of order @xmath247 or smaller . we estimates the waveform at future null infinity from data at @xmath248 . this could be made exact ( modulo the error introduced by truncating the multipole expansion ) if the nonlinear part of ( [ eq : wev ] ) were zero . thus the error introduced by ignoring @xmath241 is @xmath249 in the case of the collision of two black holes , with total mass @xmath213 and with @xmath250 , the error is @xmath251 and it is tempting to say that if extraction is performed at @xmath252 then the expected error of the we method is about 1% . this would be quite wrong because there is no reason for the constant factor in @xmath251 to be approximately @xmath253 . a numerical calculation of the emission of gravitational radiation using a ccm algorithm is expected to be second - order convergent , so that after a fixed time interval the error @xmath254 where @xmath236 is the discretization length and @xmath255 is a constant . on the other hand , the same calculation using we must allow for the error found in ( [ eq : we4 ] ) , and therefore after the same fixed time interval there will be an error of : @xmath256 where @xmath257 and @xmath258 are constants . we now estimate the amount of computation required for a given desired accuracy . we make one important assumption : * the computation involved in matching , and in waveform extraction , is an order of magnitude smaller than the computation involved in evolution , and is ignored . this is justified by the 2d nature of the extraction and matching processes as opposed to the 3d nature of evolution . for the sake of transparency we make some additional simplifying assumptions ( otherwise some extra constants of order unity would appear in the formulas below but the qualitative conclusions would be unaffected ) : 1 . the amount of computation per grid point per time - step , @xmath259 , is the same for the cauchy and characteristic algorithms . the constants @xmath260 in the equations above are approximately equal and will be written as @xmath261 . 3 . in ccm , the numbers of cauchy and characteristic grid - points are the same ; thus the total number of grid points per time - step is @xmath262 4 . in we , the outer boundary @xmath238 is at @xmath263{2 } \ , r$ ] ; thus the total number of grid points per time - step is @xmath264 it follows that the total amount of computation @xmath234 required for the two methods is : @xmath265 thus the method which requires the least amount of computation is determined by whether @xmath266 or @xmath267 . ( because of the assumptions ( 1 ) to ( 4 ) this criterion is not exact but only approximate . ) as stated earlier , the value of @xmath237 is determined by the physics , specifically by the condition that the nonlinearities outside @xmath237 must be sufficiently weak so as not to induce caustics . the value of @xmath82 is determined by the accuracy condition ( [ eq : ce2 ] ) , and also by the condition that the nonlinearities outside @xmath82 must be sufficiently weak for the existence of a perturbative expansion . thus we never expect @xmath82 to be significantly smaller than @xmath237 , and therefore the computational efficiency of a ccm algorithm is never expected to be significantly worse than that of a we algorithm . if high accuracy is required , the need for computational efficiency always favors ccm . more precisely , for a given desired error @xmath1 , eq s . ( [ eq : ce1 ] ) and ( [ eq : ce2 ] ) and assumption ( 2 ) imply @xmath268 thus @xmath269 so that @xmath270 this is the crucial result : the computational intensity of ccm relative to that of we goes to zero as the desired error @xmath1 goes to zero . the computer code described in this paper is a partial implementation of cce . that is , given data on an @xmath74 worldtube @xmath0 , the code calculates the gravitational radiation at future null infinity in the quasispherical approximation . a full implementation of cce is currently being developed which addresses the following issues : * the ignored nonlinear terms in the einstein equations must be calculated , discretized and incorporated into the code . * algorithms need to be developed to translate numerical cauchy data near @xmath0 into characteristic data on @xmath0 . * in general @xmath0 will be described in terms of cauchy coordinates , and will not be exactly @xmath271 ; the characteristic algorithm needs amendment to allow for this . once a fully nonlinear cce code has been achieved it will be possible , under certain circumstances , to obtain second - order convergence to the true analytic solution at future null infinity . for example , if @xmath0 has radius @xmath82 and the radius of the cauchy domain is @xmath238 ( @xmath272 ) , then causality implies that the gravitational field at @xmath0 will not be contaminated by boundary errors until time @xmath273 , where @xmath274 at the start of the simulation . there is no analytic error in the characteristic computation , so there will be no analytic error in the gravitational radiation at future null infinity for the initial time period @xmath275 ; under some circumstances this may be the time period that is physically interesting . further , this time period @xmath275 may be extended by using results from the characteristic computation to provide the outer boundary condition in the cauchy calculation . this would amount to a partial implementation of ccm since there would be data flow in both cauchy to characteristic , and characteristic to cauchy , directions ( the implementation is only partial because @xmath82 and @xmath238 are very different ) . since the data flow is two - way , the possibility of a numerical instability arises . however , the timescale of the growth of any instability would be @xmath275 , and therefore such a computation could be safely run for a time of several @xmath275 ; the results obtained would be second - order convergent to the true analytic solution . once the technology for cauchy to characteristic , and characteristic to cauchy , data flow across an arbitrary worldtube has been developed , a full implementation of ccm will amount to taking the limit in which the outer boundary approaches the extraction worldtube . we are encouraged to believe that this is feasible , i.e. without numerical instability , because ccm _ has _ been achieved for the model problem of the nonlinear 3d scalar wave equation @xcite . this work was supported by the binary black hole grand challenge alliance , nsf phy / asc 9318152 ( arpa supplemented ) , and by nsf phy 9510895 to the university of pittsburgh . thanks the south african frd for financial support , and the university of pittsburgh for hospitality during a sabbatical . computer time has been provided by the pittsburgh supercomputing center under grant phy860023p and by the high performance computing facility of the university of texas at austin . given a metric in a bondi null coordinate system , the gauge freedom is @xmath276 subject to the conditions @xmath277 , @xmath278 and @xmath279 . these latter conditions imply the functional dependencies @xmath280 @xmath281 and @xmath282 for a spherically symmetric background metric we drop quadratically asymmetric terms to obtain @xmath283 and @xmath284,\ ] ] where @xmath285 and @xmath286 , in terms of a complex scalar field @xmath287 . this gives rise to the following gauge freedom in the metric quantities : @xmath288 @xmath289 @xmath290 and @xmath291 we present a 3d generalization of a scheme @xcite for generating linearized solutions off a minkowski background in terms of spin - weight 0 quantities @xmath292 and @xmath293 , related to @xmath31 and @xmath63 by @xmath294 and @xmath295 . we may in this approximation choose a gauge in which @xmath296 or otherwise use the gauge freedom to set @xmath297 . in either case , @xmath66 is given by the radial integration of the linearization of ( [ eq : ww ] ) and the remaining linearized equations reduce to @xmath298 and @xmath299 where @xmath300 is the angular momentum operator . now set @xmath301 and @xmath302 then @xmath303 where @xmath304 is the wave operator @xmath305 it follows that @xmath306 suppose now that @xmath307 is a complex solution of the wave equation @xmath308 . then eq . ( [ eq;hz ] ) is satisfied as a result of ( [ eq : alpha ] ) and ( [ eq : z])and ( [ eq : ecr ] ) implies @xmath309 . if @xmath307 is smooth and @xmath310 at the origin , this implies @xmath311 , so that the linearized equations are satisfied globally . the condition that @xmath312 eliminates fields with only monopole and dipole dependence so that it does not restrict the generality of the spin - weight 2 function @xmath31 obtained . any global , asymptotically flat linearized solution may be generated this way . alternatively , given a wave solution @xmath307 with possible singularities inside some worldtube , say @xmath248 , we may generate an exterior solution , corresponding to radiation produced by sources within the worldtube , by requiring @xmath313 or @xmath314 this is a constraint on the integration constants obtained in integrating ( [ eq : alpha ] ) and ( [ eq : z ] ) which may be satisfied by taking @xmath315 and @xmath316 this determines an exterior solution in a gauge such that @xmath317 . in the characteristic formulation , the linearized equations form the principle part of the full system of bondi equations . therefore insight into the stability properties of the full evolution algorithm may be obtained at the linearized level . here we sketch the von neumann stability analysis of the algorithm for the linearized bondi equations , generalizing a previous treatment given for the axisymmetric case . the analysis is based up freezing the explicit functions of @xmath11 and stereographic coordinate @xmath318 that appear in the equations , so that it is only valid locally for grid sizes satisfying @xmath319 and @xmath320 . however , as is usually the case , the results are quite indicative of the stability of the actual global behavior of the code . setting @xmath321 and @xmath322 and freezing the explicit factors of @xmath11 and @xmath318 at @xmath248 and @xmath323 , the linearization of the bondi equations ( [ eq : wq ] ) , ( [ eq : wu ] ) and ( [ eq : wev ] ) takes the form @xmath324 and @xmath325 writing @xmath326 , introducing the fourier modes @xmath327 ( with real @xmath261 , @xmath328 and @xmath329 ) and setting @xmath330 , these equations imply @xmath331\ ] ] and @xmath332,\ ] ] representing damped quasinormal modes . consider now the fde obtained by putting @xmath333 on the grid points @xmath334 and @xmath0 on the staggered points @xmath335 , while using the same stereographic grid @xmath336 and time grid @xmath337 . let @xmath145 , @xmath62 , @xmath82 and @xmath146 be the corner points of the null parallelogram algorithm , placed so that @xmath145 and @xmath62 are at level @xmath338 , @xmath82 and @xmath146 are at level @xmath339 , and so that the line @xmath340 is centered about @xmath334 and @xmath341 is centered about @xmath342 . then , using linear interpolation and centered derivatives and integrals , the null parallelogram algorithm for the frozen version of the linearized equations leads to the fde s ( all at the same time level ) and @xmath344 where @xmath345 represents a centered first derivative . again setting @xmath330 and introducing the discretized fourier modes @xmath346 , we have @xmath347 and @xmath348 , where @xmath349 $ ] , and ( [ eq : fgamma ] ) and ( [ eq : fbigg ] ) reduce to @xmath350 = l \left [ i { r \over { \delta r } } \sin \left ( \frac{\alpha}{2 } \right ) - \frac{1}{2 } \cos \left ( \frac{\alpha}{2 } \right ) \right ] \label{eq : ffgamma}\ ] ] and @xmath351 where @xmath352 , @xmath353 and @xmath354 . the stability condition that @xmath355 then reduces to @xmath356\ge 0 $ ] . it is easy to check that this is automatically satisfied . as a result , local stability analysis places no constraints on the algorithm . it may seem surprising that no analogue of a courant - friedrichs - levy ( cfl ) condition arises in this analysis . this can be understood in the following vein . the local structure of the code is implicit , since it involves 3 points at the upper time level . the stability of an implicit algorithm does not necessarily require a cfl condition . however , the algorithm is globally explicit in the way that evolution proceeds by an outward radial march from the origin . it is this feature that necessitates a cfl condition in order to make the numerical and physical domains of dependence consistent . in practice the code is unstable unless the domain of dependence determined by the characteristics is contained in the numerical domain of dependence . it is important to note that if @xmath63 ( or @xmath0 ) are not discretized on a staggered grid then the above analysis shows the resulting algorithm to be unconditionally unstable regardless of any cfl condition .	we treat the calculation of gravitational radiation using the mixed timelike - null initial value formulation of general relativity . the determination of an exterior radiative solution is based on boundary values on a timelike worldtube @xmath0 and on characteristic data on an outgoing null cone emanating from an initial cross - section of @xmath0 . we present the details of a 3-dimensional computational algorithm which evolves this initial data on a numerical grid , which is compactified to include future null infinity as finite grid points . a code implementing this algorithm is calibrated in the quasispherical regime . we consider the application of this procedure to the extraction of waveforms at infinity from an interior cauchy evolution , which provides the boundary data on @xmath0 . this is a first step towards cauchy - characteristic matching in which the data flow at the boundary @xmath0 is two - way , with the cauchy and characteristic computations providing exact boundary values for each other . we describe strategies for implementing matching and show that for small target error it is much more computationally efficient than alternative methods .
in the local universe , three types of bulges have been defined ( see * ? ? ? * ; * ? ? ? * for recent reviews ) . _ classical bulges _ are dynamically hot systems with red colors , high concentration , and formation processes perhaps similar to elliptical galaxies . _ pseudobulges _ @xcite have lower central densities and bluer colors than classical bulges , often with ongoing star formation . they are believed to form through secular processes that drive gas and stars from the disk inward . _ boxy - peanut shaped bulges _ , sometimes considered as a subclass of pseudobulges , are formed via the natural evolution of barred galaxies . the ideas on the formation histories of bulges can be confronted with direct observations of high - redshift bulge precursors . of critical importance is the search for of strong star formation episodes in the nuclei of galaxies which would qualify as late bulge formation . recent works provide differing accounts of the abundance of blue bulges . @xcite argue for a @xmath6 of young field spheroids in the two _ hubble deep fields _ ( hdf ) . in contrast , @xcite found that @xmath7 of luminous field bulges at @xmath1 in the groth strip are nearly as red as local bulges , arguing for very little growth of massive bulges at @xmath8 . @xcite , working with goods - n field disk galaxies , noted that bulge color is related to mass : the most massive bulges are as red as massive spheroids , perhaps from mass assembly at high redshift ; smaller bulges show more diverse star formation histories , with significant star formation at @xmath8 . in our own analysis of nuclear colors of disk galaxies at @xmath9 in the groth strip ( * ? ? ? * db08 ) , we found that @xmath10 of @xmath9 bulges define a passive evolution red sequence ( rs ) ; the remaining @xmath11 have bluer colors that may trace ongoing bulge formation / growth . for such late - forming bulges two questions arise . first , what type of @xmath12 bulge ( classical , pseudobulge ) may be the descendant of an observed star - forming bulge at intermediate @xmath5 . second , who are their precursors . these may be high-@xmath5 rs classical bulges rejuvenated due to e.g. , gas infall and ensuing star formation . rejuvenation models have often been advocated ( e.g. * ? ? ? * ; * ? ? ? * ) . alternatively , blue bulges may be the result of centrally concentrated star formation in the disk , perhaps through redistribution of gas through secular or environmental processes . distinguishing rejuvenated bulges from bona - fide young bulges growing from their host disks can be done by measuring the relationship between central surface brightness and color . for the general galaxy population in the local universe , color is strongly correlated with surface density , in the sense that redder galaxies are denser @xcite , suggesting that galaxies with high surface densities formed the bulk of their stars at earlier epochs than galaxies with lower surface densities . for galaxies with bulges , @xcite conclude that galaxies with classical bulges have red global colors and high central surface brightness , whereas galaxies harbouring pseudobulges show bluer global colors and moderate central surface brightness . in this letter we present the redshift evolution of the relationship between bulge colors and central surface brightness , @xmath0 . we work with a diameter - limited sample in the redshift range @xmath13 . we quantify the @xmath5-evolution of the @xmath0 distribution for bulge and non - bulge galaxies , and investigate the ability of color-@xmath0 diagram to segregate different types of bulges at @xmath14 . our main result is that , beyond @xmath15 , blue nuclei appear at increasingly high @xmath0 , tracing episodes of nuclear star formation that lead to bulge growth , most probably by secular processes . their densities make them candidate precursors of today s pseudobulges , we do not find evidences of rejuvenation of clasical bulges . interestingly , the agn fraction increases as well among bulge galaxies at @xmath16 . a cosmology with @xmath17 , @xmath18 , @xmath19 is assumed througout . magnitudes are expressed in the vega system . .bulge sample characteristics [ cols="<,^,^,^,^ " , ] tnote.- ( 1 ) number of galaxies without bulge and without agn at each redshift range ; ( 2 ) and ( 3 ) median observed @xmath20-band @xmath0 and standard deviation . sample selection followed identical procedures as in db08 , and we refer the reader to @xcite for a detailed description , and to @xcite for details specific of the present sample . in brief , samples were selected from the acs / hst goods - n @xcite , using catalog and image versions v1.1 and v1.0 , respectively . the images cover 160 arcmin@xmath21 at 0.03 arcsec / pixel through multi - drizzle . we started from a diameter - limited parent sample with sextractor radius above 1.4 ( 398 galaxies ) . all have spectroscopic redshifts @xcite , in the range @xmath13 . we selected a subsample of galaxies with prominent bulges using the standard _ mark the disk _ method of assuming an exponential profile for the disk . we define a bulge prominence index @xmath22 that measures the excess central surface brightness over the inward extrapolation of the exponential profile of the disk ; we take as galaxies with measurable bulges those with @xmath23 magarcsec@xmath24 . the subsample of galaxies with measurable bulges comprises 131 galaxies , after excluding 53 strongly - distorted merger candidates . the remainder 214 galaxies , classified as disk galaxies without bulge , constitute a comparison sample . galaxies in the latter sample may contain small bulges , unresolved by our images . indeed , our bulge sample comprises physically large bulges , though not restricted to classical bulges ; e.g. , simulations show that the prototype peanut - shaped bulge of ngc 5965 would be clearly distinguished out to @xmath14 in our survey . see db08 , 3 , for a detailed analysis of selection effects in our bulge sample . to investigate selection biases in our diameter - limited sample , we have additionally selected samples by physical radius ( above 10.8 kpc ) ; we will show that their behavior is very similar to the angular - limited ones . the central surface brightness , @xmath0 in rest - frame @xmath20 , was measured on a central aperture with a physical diameter of 1 kpc , which corresponds to the size of the image psf for the most distant galaxy in the sample . we tested deconvolution , and bulge - disk decomposition , but both provided dubious results . we calculated @xmath0 using the nearest observed band to the rest - frame @xmath20 one at each redshift ; it was corrected of the term @xmath25 of the cosmological dimming , and of inclination and dust , the latter following @xcite and @xcite . nuclear colors were extracted from the same 1 kpc central aperture where we measured @xmath0 . integrated galaxy colors were obtained using mag_best sextractor magnitudes . for galaxies harbouring a broad - line or a narrow - line agn , nuclear colors were measured just in the origin of the color profile , and global colors excluded an inner region of radius 2 kpc , to avoid agn contamination . rest - frame @xmath2 colors were measured using the nearest observed colors at each redshift . the corrections for filter differences were calculated using seds obtained from the best - fit solution , delivered by hyperz @xcite , to photometric points in the bands @xmath26 . in fig . [ fig : magsupcenaper ] , first and second rows , we plot galaxy nuclear and global rest - frame @xmath2 colors vs @xmath0 , for three redshift ranges : @xmath27 , @xmath28 and @xmath3 . the horizontal solid and dotted lines trace the mean color of the rs and its blue boundary , respectively , at each redshift bin , from @xcite . the dashed lines trace linear least - squares fits to the color-@xmath0 distribution of bulge galaxies in the @xmath29 bin ( @xmath30 for nuclear colors ; @xmath31 for global colors ) ; these lines are also plotted in the corresponding higher @xmath5 bins for reference . the dot - dashed lines trace fits to the bulge galaxy colors in the @xmath28 range ( @xmath32 for nuclear colors ; @xmath33 for global colors ) . in the range @xmath27 , a strong correlation exists between the @xmath2 color and @xmath0 ; both nuclear and global colors are redder in galaxies with higher @xmath0 . galaxies without bulges are confined to the blue , faint region of the color-@xmath0 diagram , whereas those with bulges cover the entire width of the distribution , with a concentration at the red , high-@xmath0 end . the trend is very strong ( spearman rank correlation coefficient @xmath34 , @xmath35 for nuclear colors of bulge galaxies ; @xmath36 , @xmath37 for global colors ) with one single deviant point , a galaxy harbouring a broad - line agn . trends are similarly strong ( @xmath38 , @xmath39 for nuclear colors of bulge galaxies ) when the sample is limited by physical diameter ( [ sec : data ] ) . this shows that the color-@xmath0 correlation is not driven by large galaxies at the bright end , and small ones at the faint end . because red stellar populations , either from old age or from dust , are dimmer than younger populations , the color-@xmath0 trend implies that redder bulges have higher central stellar surface mass densities . hence , at @xmath40 , central surface density correlates with color , and , hence , with the star formation history of the bulge : denser bulges stopped forming stars in earlier epochs than lower density bulges . not only nuclear colors , but also global colors correlate with central surface density ( fig . [ fig : magsupcenaper]d ) , as found in the local universe @xcite . interestingly , @xmath0 for galaxies _ without _ bulges also correlate with nuclear colors ( @xmath41 , @xmath42 ) and with global colors ( @xmath43 , @xmath44 ) , i.e. , the scaling of central surface brightness and color is not entirely due to the presence of the bulge . if the two extremes of the color-@xmath0 distribution trace classical and pseudobulges @xcite , then fig . [ fig : magsupcenaper]a indicates that the two types of bulges are found up to @xmath45 . bluer bulges occupy the color-@xmath0 locus of star - forming galaxies , and must be forming stars . notably , all of such galaxies , presenting lower @xmath0 than their redder counterparts , must have lower surface densities too . it is unlikely that such blue bulges are rejuvenated classical bulges . if they were , the high density of the underlying classical bulge , plus the high brightness of the burst , would lead to higher surface brightness than red , classical bulges , contrary to what is observed . we quantify these statements below . at the intermediate redshift bin @xmath28 ( fig . [ fig : magsupcenaper]b , e ) , despite a brighter @xmath0 cutoff due to cosmological dimming , the color-@xmath0 relation remains in place ( @xmath46 $ ] , @xmath47 $ ] for nuclear colors of bulge galaxies [ selected by physical radius ] , and @xmath48 , @xmath49 for global colors ) . linear fits to the nuclear and to the global color-@xmath0 distributions yield very similar results to those at @xmath40 . however , blue deviant nuclei are found around @xmath50 and @xmath51 ( fig . [ fig : magsupcenaper]b ) . the color profiles of those objects ( not shown ) shows mild blueing toward the center , which provides evidence for enhanced star formation in the nuclear region of the galaxy at redshifts @xmath28 . two of the four blue deviant bulges may be minor merger remnants , and one hosts a bar , which suggests that these bulges might be growing by secular processes caused by disk instabilities or by satellite accretion @xcite . like in the lower redshift bin , we do not find evidence for rejuvenation of red , classical bulges at @xmath28 : the three blue , high - surface brightness nuclei in figure [ fig : magsupcenaper]b are agn hosts . finally , at @xmath3 , while global colors still correlate with @xmath0 ( fig . [ fig : magsupcenaper]f : @xmath52 , @xmath53 ) , nuclear colors ( fig . [ fig : magsupcenaper]c ) show a null trend ( @xmath54 , @xmath55 ) . figure [ fig : magsupcenaper]c shows a population of blue , high surface brightness nuclei with @xmath56 , i.e. , as high as the red , classical bulges . in the previous paragraph we already described blue , deviant nuclei at @xmath28 , but of a lower @xmath57 . the blue colors and high surface brightness must trace strong episodes of star formation concentrated in the galaxy central regions . the phenomenon may be related to the hdf - n field spheroidals with blue cores @xcite or to the luminous compact blue galaxies in the hdf - ff @xcite . one of the blue nuclei harbours a narrow - line agn ; the others are undoubtedly stellar . examples ( non - agn cases ) are shown in figure [ fig : morphology ] . one shows signs of interaction ; the others are normal spiral galaxies . the color profiles reveal strongly inverted color profiles , bluer inward , in the inner @xmath58 kpc . selection effects also contribute against a color-@xmath0 correlation in @xmath3 . missing small galaxies , due to our apparent diameter selection , is probably unimportant : the trends and correlation coefficients in figures [ fig : magsupcenaper]a - c are similar when the samples are selected by physical diameter @xmath59 kpc . but galaxies with small _ bulges _ may have been classified in the bulge - less class due to resolution and dimming effects ( db08 , 3 ) . hence , it is plausible that the color-@xmath0 correlation survives at @xmath3 if small bulges do exist at those redshifts . in any case , the @xmath3 color-@xmath0 distribution remains peculiar for the abundance of blue nuclei as bright as the classical bulges : @xmath60 of the @xmath61 nuclei have colors below @xmath62 . figure [ fig : magsupcenaper ] clarifies the redshift evolution of the color-@xmath0 distribution . the linear fits to the bulges , plotted in the diagrams , show that there is no significant change between the ranges @xmath40 and @xmath28 . but for @xmath3 , the distribution lies well below the fit to the first range , with brighter @xmath0 and bluer colors . the variation in surface brightness is shown in the @xmath0 histograms ( fig . [ fig : magsupcenaper]ghi ) : the peaks of the distributions of both bulge and non - bulge galaxy samples shift toward brigther @xmath0 , by @xmath63 mag , from @xmath64 to @xmath65 ( see tables [ tab : medians ] and [ tab : mediansnb ] ; median values do not change significantly if we impose a common selection limit at all @xmath5 , @xmath66 $ ] for the bulge [ non - bulge ] sample ) : both galaxies with and without bulges were brighter and bluer at @xmath14 and faded and reddened , toward @xmath64 . most likely , the centrally - concentrated star forming events contribute to bulge formation inside disks . who are the precursors , and who are the descendants of such blue , high - surface brightness nuclei ? in particular , are they rejuvenated classical bulges that will evolve back into the rs once the burst has faded ? we estimate the evolution of starbursting nuclei in color-@xmath0 space using composite stellar population ( csp ) models from ( * ? ? ? * bc03 ) , with exponentially - decaying star formation rates ( sfr ) , without dust , and build tracks intended to match the span of color-@xmath0 at @xmath64 . for the reddest track , we look for models with @xmath67 , and assign them a surface brightness of @xmath68 . color and surface brightness are tracked back in time using bc03 and the variation with @xmath5 of the angular size of the galaxy . for a canonical @xmath69 , and a moderate @xmath70 gyr , the model evolves along the orange lines in fig . [ fig : magsupcenaper]abc , becoming both bluer and brighter as we move back in time . at @xmath3 , the evolution fails to reach the region of the blue , high - surface brightness colors of the starbursting nuclei seen at those redshifts . the bluer colors at @xmath1 require a younger age , but a younger age would yield too high a surface brightness . indeed , evolution is not vertical but diagonal in the color-@xmath0 plane , as population ageing implies both reddening and dimming . taking into account that starbursting nuclei are likely composite old plus young populations does not help , as the underlying , more stable old population would contribute to the surface brightness at all ages , making the problem worse . hence , the track shows that the blue , high - surface brightness nuclei seen at @xmath3 are not descendants , nor precursors , of the red , high - surface brightness classical bulges . for the bluest track , we impose a @xmath12 color of @xmath71 , and assign a @xmath12 surface brightness @xmath72 . this track ( green line on fig . [ fig : magsupcenaper]abc ) was modeled using a @xmath73 and a @xmath74 gyr . blue deviant nuclei with intermediate surface brightness at @xmath28 fit the track evolution . an intermediate model with a slightly brighter @xmath0 at @xmath64 , @xmath75 , and a @xmath76 ( violet line on fig . [ fig : magsupcenaper]abc ) , nicely fits the blue , high - surface brightness nuclei at @xmath3 . this model links the blue , high-@xmath0 deviant bulges at @xmath3 with moderate color , intermediate-@xmath0 bulges at @xmath12 ; it supports the idea that these bulges are precursors of local universe pseudobulges , rather than old , dense bulges . classifying as agn those galaxies with @xmath77 ergs@xmath78 in the 28 kev band of _ chandra _ @xcite , @xmath79 of the galaxies with ( without ) bulges host an agn in the @xmath80 range . this highlights that the preference of agn to live in galaxies with bulges extends up to @xmath65 . in fig . [ fig : magsupcenaper ] we denote with specific symbols three agn types , namely broad - line ( bl ) , narrow - line ( nl ) , and low - luminosity ( ll ; no agn spectral lines ) , from @xcite . except for the three bl agn , the nuclear colors and @xmath0 of galaxies with agn are in the general region occupied by the non - agn galaxies both in the main color-@xmath0 sequence , and in the blue , high-@xmath0 region of the starbursting nuclei . their nuclear colors are likely to be affected by the agn , especially for the more x - luminous ( nuclear colors for our agn galaxies get bluer with increasing x - ray luminosity ) . however , the size of the blue nuclear region , and the comparison with synthetic color profiles built with combinations of a disk , a bulge , and a nuclear source with varying colors and luminosity ratios , do suggest that most blue regions are spatially resolved , hence that star formation coexists with the agn . the trigger of the agn activity in these galaxies is unclear . morphologies for the three bl agn hosts reveal a normal , bulge - dominated spiral structure , with a central bar in two of them . the nl agn galaxies show a more mixed variety of types , including early - type galaxies , normal and barred spirals , and distorted galaxies . hence , agn activity is not associated with merging or interacting situations , as noted by @xcite . in the goods - n field , @xmath60 of very blue , high-@xmath0 bulges in @xmath3 trace an epoch of important bulge formation inside disks . the fraction of blue cores is close to that for hdf field spheroidals @xcite . similar but weaker nuclear star formation is present at @xmath28 , pointing at bulge - forming events that become weaker with cosmic time , and lead to a mass - age relation for bulges @xcite . the data imply important bulge growth at @xmath8 . models predict that such starbursting nuclei evolve into moderate surface - brightness , intermediate color @xmath12 pseudobulges rather than classical bulges , and argue against rejuvenation processes for @xmath14 dense and old bulges . the latter are already abundant at @xmath3 , as found by @xcite . we thank the anonymous referee for suggestions that improved the paper and ignacio trujillo and mercedes prieto for useful discussions . this work is based on archival data from the hst , obtained from the data archive at stsci , which is operated by aura under nasa contract nas5 - 26555 . support has been provided by the spanish programa nacional de astronoma y astrofsica through project number aya2006 - 12955 and through the consolider - ingenio 2010 program grant csd2006 - 00070 : first science with the gtc ( http://www.iac.es/consolider-ingenio-gtc/ ) . , d. m. , bauer , f. e. , brandt , w. n. , schneider , d. p. , hornschemeier , a. e. , vignali , c. , barger , a. j. , broos , p. s. , cowie , l. l. , garmire , g. p. , townsley , l. k. , bautz , m. w. , chartas , g. , & sargent , w. l. w. 2003 , , 126 , 539 , m. , ferguson , h. c. , koekemoer , a. m. , dickinson , m. , alexander , d. m. , bauer , f. e. , bergeron , j. , biagetti , c. , brandt , w. n. , casertano , s. , cesarsky , c. , chatzichristou , e. , conselice , c. , cristiani , s. , da costa , l. , dahlen , t. , de mello , d. , eisenhardt , p. , erben , t. , fall , s. m. , fassnacht , c. , fosbury , r. , fruchter , a. , gardner , j. p. , grogin , n. , hook , r. n. , hornschemeier , a. e. , idzi , r. , jogee , s. , kretchmer , c. , laidler , v. , lee , k. s. , livio , m. , lucas , r. , madau , p. , mobasher , b. , moustakas , l. a. , nonino , m. , padovani , p. , papovich , c. , park , y. , ravindranath , s. , renzini , a. , richardson , m. , riess , a. , rosati , p. , schirmer , m. , schreier , e. , somerville , r. s. , spinrad , h. , stern , d. , stiavelli , m. , strolger , l. , urry , c. m. , vandame , b. , williams , r. , & wolf , c. 2004 , , 600 , l93 , n. a. , conselice , c. j. , chatzichristou , e. , alexander , d. m. , bauer , f. e. , hornschemeier , a. e. , jogee , s. , koekemoer , a. m. , laidler , v. g. , livio , m. , lucas , r. a. , paolillo , m. , ravindranath , s. , schreier , e. j. , simmons , b. d. , & urry , c. m. 2005 , , 627 , l97 , d. c. , simard , l. , willmer , c. n. a. , gebhardt , k. , bouwens , r. j. , kauffmann , g. , crosby , t. , faber , s. m. , harker , j. , sarajedini , v. l. , vogt , n. p. , weiner , b. j. , phillips , a. j. , i m , m. , & wu , k. l. 2005 , , 157 , 175 , j. 1993 , in international astronomical union . symposium , kluwer academic publishers , vol . 153 , galactic bulges : proceedings of the 153rd symposium of the international astronomical union , ed . h. dejonghe & h. j. habing , 209+ , g. d. , willmer , c. n. a. , amico , p. , chaffee , f. h. , goodrich , r. w. , kwok , s. , lyke , j. e. , mader , j. a. , tran , h. d. , barger , a. j. , cowie , l. l. , capak , p. , coil , a. l. , cooper , m. c. , conrad , a. , davis , m. , faber , s. m. , hu , e. m. , koo , d. c. , le mignant , d. , newman , j. a. , & songaila , a. 2004 , , 127 , 3121	we analyze central surface brightness @xmath0 , nuclear and global colors of intermediate redshift disk galaxies . on an apparent - diameter limited sample of 398 galaxies from acs / hst _ great observatories origins deep survey north _ ( goods - n ) , we find 131 galaxies with bulges and 214 without . up to @xmath1 , blue , star - forming nuclei are found in galaxies with low @xmath0 only ; all high-@xmath0 nuclei show red , passive colors , so that nuclear and global @xmath2 colors strongly correlate with central surface brightness , as found in the local universe . at @xmath3 , a fraction @xmath427% of the high - surface brightness nuclei show blue colors , and positive nuclear color gradients . the associated nuclear star formation must lead to bulge growth inside disks . population modeling suggests that such blue bulges evolve into local pseudobulges rather than classical bulges . we do not find evidence for rejuvenation of classical bulges at the sampled @xmath5 . high luminosity agns become common at @xmath3 , perhaps pointing to a role of agn in the growth or star formation truncation of bulges .
understanding the formation and evolution of early - type galaxies is a key issue when addressing the mass assembly and star formation history of the galaxy population as a whole and the formation of structure in the universe , as 50% or more of all stars in the present day universe are in early - type galaxies and bulges ( see , e.g. , bell et al . 2003 ) . in hierarchical galaxy formation theories ( e.g. , cole et al . 2000 ) , massive galaxies assemble late , such that strong evolution of the mass density from @xmath8 to the present day is expected ( see , e.g. , kauffmann & charlot 1998 ) . measuring the mass density requires a measurement of the luminosity density , and an accurate determination of the @xmath2 . @xmath2 can be estimated from models ( see , e.g. , bell et al . 2004 ) , but these estimates are uncertain due to the age / metallicity degeneracy and the unknown imf of the stellar populations of the galaxies ( bruzual & charlot 2003 ) . the fundamental plane ( djorgovski & davis 1987 ; dressler et al . 1987 ) provides a tool to measure the evolution of @xmath2 without model uncertainties . the @xmath2 offset of high redshift galaxies from the local fp can be used to calibrate high redshift galaxy masses and to estimate the age of their stellar populations ( franx 1993 ) . this technique has been used successfully to measure the luminosity weighted ages of massive cluster galaxies , which have formed most of their stars at redshifts @xmath9 ( see , e.g. , van dokkum & franx 1996 ; van dokkum & stanford 2003 , holden et al . 2005 ) . however , it is not clear whether galaxies in the general field evolve in the same way . in fact , in the hierarchical picture the formation redshift of galaxies with a given mass depends on environment ( diaferio et al . this would lead to substantial age differences between field and cluster galaxies at any redshift ( van dokkum et al . since this is a generic property of all hierarchical formation models , measuring this difference is a critical test for those theories . various authors have measured the @xmath2 evolution of field early- type galaxies through deep spectroscopy of magnitude limited samples . the results are much less conclusive than the results from cluster studies and the comparison between field and cluster has proved to be very hard . some authors claim much faster evolution for field galaxies than for cluster galaxies ( treu et al . 2002 ; gebhardt et al . 2003 ) , but others find that field and cluster galaxies evolve at comparable rates ( van dokkum et al . 2001 , van dokkum & ellis 2003 , van der wel et al . 2005 ) . studies involving lensing galaxies ( kochanek et al . 2000 , rusin et al . 2003 , van de ven et al . 2003 ) , indicate the presence of a mix of fast and slowly evolving galaxies . it is unclear whether the differences between the various results are caused by selection effects , measurement errors due to low signal - to - noise spectra , low number statistics , or contamination by late - type galaxies . this paper describes a study of early - type galaxies at @xmath10 using much higher quality data than in previous studies . the substantially large number of objects with very high signal - to - noise spectra enables us to accurately measure the @xmath2 evolution of the field early - type galaxy population , to compare the cluster and field populations , to study correlations between @xmath2 , @xmath11 , and rest - frame color , and to describe the possible effects of biases . also , we carefully compare the samples and results from previous studies and this study in order to verify previous claims about the evolution of field galaxies , and to see whether previous results are in fact consistent with each other and these new results . in section 2 we describe the sample selection , the spectroscopic observations and data reduction , and the measurement of velocity dispersions . section 3 describes the measurement of the structural parameters , colors , morphologies , and the available x - ray data . in section 4 we present the results throughout this paper we use vega magnitudes , and assume @xmath12 , with a hubble constant of @xmath13 km s@xmath14 mpc@xmath14 . we selected galaxies in the chandra deep field - south ( cdfs ) and the rdcs1252.9 - 2927 cluster field ( cl1252 , rosati et al . 2004 ) , which both have deep optical imaging from the advanced camera for surveys ( acs ) on the hst . goods provides publicly available imaging in four filters ( giavalisco et al . 2004 ) : f475w , f606w , f775w , and f850lp ( hereafter @xmath15 , @xmath16 , @xmath17 , and @xmath18 ) . as these data were not yet available when we started this project , we used ground - based combo-17 photometry ( wolf et al . 2004 ) to select our sample for the first observing run . for subsequent runs version 0.5 of the acs goods data were available , and for the last run we used the version 1.0 data release . ( blakeslee et al . 2003 ) provide acs imaging on the cl1252 field in the @xmath17 and @xmath18 bands . in order to construct a sample of early - type galaxies at @xmath10 in the cdfs , we selected objects with @xmath19 and combo-17 photometric redshifts in the range @xmath20 . ( we use @xmath18 when we mean redshift , and @xmath21 if we mean @xmath18-band magnitude , but when indicating a color we omit the @xmath22 subscript for clearity . ) this color cut selects galaxies redder than a local sbc galaxy at @xmath8 . therefore , this study only includes galaxies that are on the red sequence at @xmath10 . we morphologically classified all galaxies satisfying these criteria and brighter than @xmath23 , distinguishing between early- and late - type galaxies using the acs imaging . the classification was based on compactness , regularity and the presence of spiral arms . 26 out of the 52 galaxies satisfying our selection criteria we classified as early - type galaxies . we designed multi - slit masks for three different pointings , selected by the number of primary targets that could be included . open spaces in the masks were filled with early - type galaxies satisfying the color and redshift criteria but fainter than @xmath23 , early - type galaxies with lower photometric redshifts and late - type galaxies with @xmath24 and redshifts @xmath20 . switching from using ground - based @xmath25 colors from combo-17 to acs @xmath25 colors from goods did not lead to large differences between the selected samples , although several objects changed priority . the same selection criteria were used for the cl1252 field . however , the mask design for the cl1252 field was geometrically constrained because the primary targets were cluster galaxies at @xmath26 . the two brightest galaxies in the cluster are the two central galaxies , which had to be included in a single slit because of their small angular separation of @xmath27 . therefore , not only the positions , but also the position angles of the designed masks were fixed . unfortunately , only two galaxies brighter than @xmath23 , redder than @xmath28 and early - type morphologies could be included , additionally to the cluster galaxies . similarly to the cdfs masks , fillers were included . we carried out the observations with fors2 in mxu mode on eso s kueyen , one of the vlt unit telescopes . we used the 600z grism together with the og590 order separation filter to obtain a sufficiently high spectral resolution ( @xmath29 km s@xmath14 ) and to cover the wavelength range around the balmer/4000 break for galaxies at @xmath10 . the observations were carried out in series of four dithered exposures with spatial offsets of @xmath27 or @xmath30 and equal exposure times ranging from 14 to 30 minutes each . in total , 51 hours of scientifically useful integration time was acquired , of which 38 hours had seeing better than @xmath31 . the cumulative integration time for the three pointings in the cdfs is 27 hours , with a median seeing of @xmath32 . the single pointing in the cl1252 field has an integration time of 24 hours , with a median seeing of @xmath33 . these observations were carried out during five different observing runs from september 2002 to november 2003 . the sample described in this paper consists of 38 galaxies with velocity dispersions , of which 20 are early - type galaxies at @xmath10 , and 18 are early - type galaxies at lower redshift , or late - type galaxies . 100% of our primary targets yielded velocity dispersions . the cl1252 observations also yielded four velocity dispersions of cluster galaxies . the fp of the 1252 cluster is discussed by ( holden et al . 2005 ) . the spectroscopic data were reduced using standard iraf tasks . lamp flat fields were taken before or after each night , in sequences of five exposures . we used the sequence closest in time to the science observation . cosmic rays were removed using the l.a.cosmic task ( van dokkum 2001 ) . afterward , all frames were checked manually . we subtracted a two - dimensional sky spectrum from each exposure , obtained by median averaging the four dithered exposures in a sequence , masking the target and secondary or serendipitous objects , if present . the atmospheric emission lines , which are bright and abundant in the observed wavelength range , were used to perform the wavelength calibration . we corrected for distortion in the spatial direction by tracing the target . all individual exposures were optimally weighted to obtain maximum @xmath34 . there are various atmospheric absorption features in the observed wavelength range . because the strength and shape of these features change with airmass and atmospheric conditions we needed to correct each exposure separately . to this end we included a blue star in each of our masks , which was reduced along with the galaxy spectra . after the final combination , the regions in the galaxy spectra with atmospheric features were divided by the normalized spectrum of the blue star . spectroscopic standard stars were used to do a relative flux calibration . one - dimensional spectra were extracted by adding those pixel rows with more than 25% of the flux of the brightest row , weighting optimally . the smoothed one - dimensional spectra are shown in figure [ fig : spec ] . the coordinates of the objects for which we measured velocity dispersions ( see section [ sig ] ) in table 1 . redshifts , @xmath34 and emission lines are given in table 2 . velocity dispersions are obtained by fitting template spectra to the observed galaxy spectra . the fitting method is extensively described by ( van dokkum & franx 1996 ) . the continua of both the observed and the template spectra are filtered out in fourier space and the template spectrum is convolved with a gaussian to match the width of the absorption lines in the galaxy spectrum . the part of the galaxy spectrum used in the fit is as large as possible . therefore , our measurements do not rely on a few high @xmath34 absorption features . as templates we use coud spectra of 132 stars with the appropriate wavelength range from the sample constructed by ( valdes et al . 2004 ) , with a spectral range from f0 to m6 , including both very low and high metallicity stars , and different luminosity classes . these spectra have a fwhm resolution of about 1 . each stellar spectrum needs to be smoothed to each galaxy spectrum separately before being re - binned . a second order function is fitted to the width of atmospheric emission lines as a function of wavelength to obtain the spectral resolution to which the template spectra are smoothed . when fitting the galaxy spectra , we weight with the inverse of the sky brightness , and we mask the region around the atmospheric a band at 7600 . the spectrum above 9300 is omitted because of the strong atmospheric absorption , the ever increasing brightness of the sky emission lines and the decreasing system throughput . after performing the fit for a small number of templates , masking and weighting as described , we check the residuals from the fit . regions with emission lines , large sky line residuals and remaining data artifacts such as cosmic ray remnants are masked if present . we then apply fit the galaxy spectrum with all template spectra . we check whether the obtained parameters change strongly if one or two strong features are masked out , but we conclude that this generally is not the case : excluding the strongest features from the fit increases the @xmath35-value but does not change the results significantly in most cases . for some spectra , however , including balmer lines in the fit leads to different results , probably because unseen emission line contributions contaminate these features . for low quality spectra ( @xmath36 per 1.6@xmath37 pixel in the extracted , one - dimensional spectra ) the contributions of balmer lines or other strong features can hardly be checked because excluding these strong features leaves insufficient signal to obtain a proper fit . therefore , we exclude objects with @xmath38 spectra from our analysis , but we mention the effect of including these . for all spectra with @xmath39 the random errors are below 3% for @xmath40 km s@xmath14 , and below 5% for @xmath41 km s@xmath14 . adding a systematic uncertainty ( including template mismatch and the error on the resolution of the galaxy spectra ) of about 10% for @xmath42 spectra and 2% for the highest @xmath34 spectra the total errors range from 3% to 17% with a median of 7.5% for our sample of early - type galaxies with @xmath39 . some galaxies have measured velocity dispersions that are not much larger than the resolution of the spectra . although these are included in the analysis , they play no important role in the derivation of our results . the best fitting stellar spectral type and the measured velocity dispersion are listed in table 1 . these velocity dispersions are aperture corrected to a @xmath43 diameter circular aperture at the distance of coma as described by ( jrgenson , franx , & kjrgaard 1995 ) . this correction ranges from @xmath44 to @xmath45 . the acs provides us with an unprecedented combination of deep and high resolution imaging . the spatial resolution ( fwhm ) at @xmath8 is @xmath46kpc , allowing us to accurately measure the effective radii of early - type galaxies at this redshift , which typically are a few kpc . we use the single , unstacked , flat - fielded frames publicly available through the hst mast archive . for the cdfs the number of frames for different positions ranges from 8 to 24 ( with @xmath47s exposure time each ) . for the cl1252 field the number of frames ranges from 10 to 40 ( with @xmath48s exposure time each ) , but is mostly 10 as only the center of the cluster has 40 overlapping images . for each galaxy each individual @xmath18 band image is fitted by @xmath49-models ( with @xmath50 ) convolved by a position dependent psf created with tinytim ( krist 1995 ) , measuring @xmath51 ( the effective radius ) , @xmath52 ( the surface brightness at @xmath51 ) , the position angle and the ellipticity . each individually derived set of model parameters is distortion corrected by calculating the pixel scales in the x- and y - directions , using the polynomial distortion coefficients available through the www . we then average the results and compute the measurement error from the scatter . the error is generally about 6% in @xmath51 , but the combination of the uncertainty in @xmath51 and @xmath52 is such that it is directed almost parallel to the local fp . the error relevant to the offset from the fp is typically 2% . thus , uncertainties in the offset from the local fp are dominated by the uncertainty in @xmath53 , of which the error is pointed almost perpendicular to the fp . the effective radii and surface brightnesses are given in table 1 . for consistency with earlier studies , these are the values obtained from fitting a de vaucouleur profile in our analysis . to transform the observed @xmath18 band surface brightnesses to the rest - frame @xmath54 band , we use the technique described by ( van dokkum & franx 1996 ) , using the templates from ( coleman , wu , & weedman 1980 ) and observed colors ( see section [ sec : color ] ) to interpolate between the pass - bands . the calculated rest - frame @xmath54 band surface brightnesses only depend very weakly on the spectral type of the template used . the typical difference found for using the sbc template instead of the e template is less than 0.02 mag . the transformations are a function of redshift . as an example we give the transformation ( based on the e template ) for a galaxy at @xmath8 : @xmath55 physical sizes and rest - frame @xmath54-band surface brightnesses are given in table 3 . figures [ fig : mos ] and [ fig : moslate ] show the combined residuals of the @xmath56-fits along with the color images of the 38 galaxies with velocity dispersions . figure [ fig : mos ] shows the early - type galaxies , figure [ fig : moslate ] the late - type galaxies . two numbers are used to characterize the magnitude of the residuals . at the upper right of each residual image the absolute value of the flux in the asymmetric part of the residual is given as a percentage of the total flux of the galaxy , obtained by subtracting the residual rotated by 180 degrees from the residual itself . in the upper left the absolute value of the flux in the symmetric part of the residual is given as a percentage of the total flux of the galaxy . we find that the asymmetric residual is a good indicator of morphology . our final morphological qualification is a combination of the magnitude of the asymmetric residual and visual inspection of the cause of the asymmetry . some of the early - type galaxies ( figure [ fig : mos ] ) have significant asymmetric residuals , but these are caused by features on a very small scale , in the centers of the galaxies , i.e. not by features attributed to spiral arms or other large scale irregularities . the late - type galaxies ( figure [ fig : moslate ] ) have large asymmetric residuals , caused by large scale structures like spiral arms . we note that the sercic number does not distinguish well between late- and early - type galaxies . for example , the most massive galaxy in our sample has n=2 and some galaxies that we classify as late - type galaxies have n=4 . we supplement the acs imaging with ground - based optical and near - ir imaging from fors2 and isaac on the vlt and sofi on the ntt ( vandame et al . , in preparation ) . goods acs imaging of the cdfs provides photometry in the @xmath15 , @xmath16 , @xmath17 , and @xmath18 bands ( data release version 1.0 ) , and eso s imaging survey provides sofi and isaac imaging in the @xmath57 and @xmath58 bands . since the cdfs is not entirely covered by isaac we use the sofi data for the objects outside the isaac pointings . all images were smoothed to match the resolution of the @xmath58 band data with the worst seeing , which is @xmath59 for the isaac imaging and @xmath60 for the sofi imaging . the photometric differences between the isaac and sofi datasets are small ( @xmath61 mag ) , since the zero - points of the isaac data are based on sofi photometry . for the cl1252 field we use optical imaging from acs ( @xmath17 and @xmath18 ) and fors2 ( @xmath54 , @xmath62 , and @xmath63 ) and near - ir imaging from isaac ( lidman et al . again , all images were smoothed to match the ground - based data with the worst seeing , which is @xmath64 in the @xmath54 band . we measure the flux in each band for each of our spectroscopic targets within several apertures with different radii ( @xmath65 and @xmath59 ) . contaminating objects within the aperture are masked . we choose to use the fluxes measured within a radius of @xmath66 , with a minimum of @xmath59 , trading off between the measurement accuracy and the amount of contamination . the @xmath25 and @xmath67 colors are given in table 2 , as well as total _ auto _ @xmath18 band and @xmath58 band magnitudes , as obtained with sextractor ( bertin & arnouts 1996 ) . the availability of near - ir photometry not only allows us to compute rest - frame @xmath68 colors , but also rest - frame @xmath69 colors . to transform the observed colors to rest - frame @xmath68 and @xmath69 we use the same method as used to calculate rest - frame @xmath54 band surface brightnesses ( see section [ sec : color ] ) . for @xmath8 , using the e template , the transformation is : @xmath70 @xmath71 rest - frame colors are given in table 3 . images at www.strw.leidenuniv.nl/~ vdwel / private / fppaper/ [ fig : mos ] [ fig : mos ] images at www.strw.leidenuniv.nl/~ vdwel / private / fppaper/ [ fig : moslate ] for both the cdfs and the cl1252 field deep x - ray data from chandra are available , such that we can check for the presence of agn in our galaxy sample . giacconi et al . ( 2002 ) and alexander et al . ( 2003 ) provide catalogs of the cdfs data . the chandra data of the cl1252 field is described by ( rosati et al . 2004 ) , who also constructed a point source catalog . eight galaxies in our sample of 38 are identified as agn , based on their large x - ray luminosities ( typically @xmath72erg s@xmath14 ) . five of these are early- type galaxies , of which two have emission lines in their spectra . besides the eight agn , two galaxies in our sample are identified as extended x - ray sources . this x - ray radiation is accounted for by diffuse halo gas . the x - ray luminosities of cdfs-4 and cdfs-22 are @xmath73 erg s@xmath14 and @xmath74 erg s@xmath14 , respectively . it is not surprising that cdfs-4 and cdfs-22 turn out to be two of the most massive galaxies in our sample . also , cdfs-22 is one of the galaxies with an agn . the x - ray properties of our sample of galaxies are given in table 2 . jrgensen , franx , & kjrgaard ( 1996 ) have shown that the fp for cluster galaxies in the local universe can be described by @xmath75[eq:3 ] where @xmath76 is the effective radius in kpc , @xmath53 the central velocity dispersion in km s@xmath14 , @xmath77 the surface brightness in the @xmath54 band ( @xmath78 ) , and @xmath79 the intercept . the values of the coefficients are derived from the early - type galaxies in ten nearby clusters , and imply that mass and @xmath2 scale as @xmath80 . from the sample of faber et al . ( 1989 ) we derived that the intercept of the fp lies lower by 0.04 in the projection given above . the offsets of the high-@xmath18 galaxies are computed using a local fp with coefficients from jrgensen et al . ( 1996 ) but with the intercept derived from the faber et al . ( 1989 ) sample . besides the form of the local fp , we need a large sample of local field early - type galaxies , for example to address issues such as selection effects . we construct our local field sample as in bernardi et al . ( 2003 ) : we take early - type galaxies from the sloan digital sky survey and include those with less than 10 neighbors brighter than @xmath81 and closer than @xmath82 . we convert the rest - frame @xmath83 and @xmath84 band surface brightnesses to a @xmath54 band surface brightness using the conversion given by jrgensen et al . ( 1996 ) . as we need the surface brightness in the @xmath83 and @xmath84 bands at the same radius , and bernardi et al . derive effective radii separately for each band , we compute the surface brightness in the @xmath84 band at the effective radius as measured in the @xmath83 band . we note that the fp coefficients as derived by bernardi et al . are different from those from jrgensen et al . , but this does not lead to different results . in figure [ fig : fpe ] we show the fp of the sdss galaxies and our sample , where the surface brightnesses of all galaxies are transformed to the value they would have at @xmath8 , assuming luminosity evolution found for massive cluster galaxies , @xmath85 . this value for the @xmath2 evolution of massive cluster galaxies is derived from compiling all existing data in the literature for galaxies more massive than @xmath86 ( van dokkum & franx 1996 ; kelson et al . 2000 , van dokkum & stanford 2003 ; wuyts et al . 2004 , holden et al . our field sample shown in figure [ fig : fpe ] includes all early - type galaxies with spectra with @xmath39 . this is also the sample used in the analysis throughout the rest of the paper , and in the subsequent figures . as can be seen , the fp already existed at @xmath8 for a large range in size . at low masses outliers occur , but the interpretation is not straightforward , as selection effects play a major role in this regime ( see section 4.3 ) . the offset of high redshift galaxies from the local fp is interpreted as a difference in @xmath2 as compared to equally massive local galaxies ( see , e.g. , vam dokkum & franx 1996 ) . figure [ fig : z_dml ] shows the offsets of our field galaxy sample in @xmath87 as a function of redshift . the values are listed in table 3 . cluster samples from the literature are also shown . the galaxies in our sample seem to evolve faster than the galaxies in the cluster samples , and the scatter in @xmath87 is large . before we interpret the scatter and the apparent difference between field and cluster galaxies , we need to investigate the origin of the scatter . in figure [ fig : bi_ml ] we show @xmath88 as a function of the rest - frame @xmath69 color . galaxies with low @xmath2 are bluer than galaxies with high @xmath2 , as expected from the stellar population models shown in the figure . in figure [ fig : ub_ml ] we show a similar relation between rest - frame @xmath68 and @xmath88 , but in this case the correlation is less clear , due to the fact that the range of @xmath68 colors is much smaller than the range of @xmath69 colors , and probably also because @xmath68 is more sensitive to small variations in the star formation history . considering the correlation between color and @xmath2 and the fact that @xmath2 evolves with redshift , one expects that color evolves with redshift as well . in figure [ fig : z_bi ] we show @xmath69 as a function of redshift ; note the strong similarities between figures [ fig : z_dml ] and [ fig : z_bi ] . the strong correlation between color and @xmath2 , and the similarity in @xmath2 evolution and color evolution confirm that the observed evolution and scatter of @xmath2 are intrinsic , and not due to measurement errors . we calculate the evolution of @xmath2 of our field sample by performing a linear fit and minimizing the mean deviation , weighting by the inverse of the error , and forcing the fit to go through the @xmath89 data point derived from the faber et al . ( 1989 ) sample . we separately consider the evolution of the primary sample , which contains the galaxies satisfying all of our selection criteria . we find that the average evolution of our entire early - type galaxy sample is @xmath90 ( @xmath91 for the primary sample alone ) , which is significantly faster than the evolution found for cluster galaxies , which is @xmath92 ( van dokkum & franx 1996 ; kelson et al . 2000 , van dokkum & stanford 2003 ; wuyts et al . 2004 , holden et al . 2005 ) . the scatter in @xmath87 is @xmath93 for our field galaxy sample ( 0.54 for the primary sample ) , and @xmath94 for the ms1054 cluster sample ( wuyts et al . 2004 ) . figure [ fig : fpe ] suggests that the @xmath2 evolution may depend on galaxy mass , as galaxies with small @xmath51 and low @xmath53 tend to lie lower with respect to the local fp as compared to galaxies with large @xmath51 and high @xmath53 . this was also found for cluster galaxies by wuyts et al . we estimate @xmath11 and @xmath2 in solar units as described by van dokkum & stanford ( 2003 ) . the values are listed in table 3 . we explore the mass dependence by color coding the galaxies in figure [ fig : z_dml ] according to their masses . red points are galaxies with masses larger than @xmath86 ; blue points are galaxies that have masses lower than @xmath86 . there is a striking difference between low and high mass galaxies . for galaxies with masses @xmath95 we measure @xmath96 for our field sample and @xmath97 for the cluster samples . for the massive galaxies in the primary sample alone we find @xmath98 . the observed scatter is decreased to @xmath99 for the field sample ( 0.32 for the primary sample ) , and to @xmath94 for the cluster samples . when changing this mass cut to @xmath100 , as is done by wuyts et al . ( 2004 ) , but thereby limiting the number of galaxies in our sample to four , we find @xmath101 and @xmath102 for our sample and the cluster samples , respectively . we conclude that for high - mass galaxies , there is no difference between the cluster samples and our field sample . the galaxies with masses @xmath103 in our sample evolve much faster : @xmath104 ( @xmath105 for the primary sample alone ) . we verify that these results do not change if galaxies with spectra with @xmath38 are included as well . therefore , the accuracy of our results is not limited by the quality of the velocity dispersions . in figures [ fig : z_dml ] and [ fig : z_bi ] we show evolutionary tracks for a single stellar population for formation redshifts @xmath8 and 2 . this very simple model assumes that luminosity evolves with time as @xmath106 , where @xmath107 is derived from stellar population models ( see van dokkum et al . ( 1998 ) for more details ) . these tracks indicate a large spread in formation redshifts . we note , however , that @xmath107 is sensitive to the imf . here we use a single stellar population from bruzual & charlot ( 2003 ) with a salpeter imf and solar metallicity , which yields @xmath108 in the @xmath54 band and @xmath109 in the @xmath110 band . according to this model the massive galaxies have high luminosity weighted formation redshifts ( @xmath111 ) , whereas less massive galaxies in our sample have lower formation redshifts ( @xmath112 ) . figure [ fig : m_ml ] illustrates the tight relation between mass and @xmath88 . we also show the sdss field sample described in section 4.1 . the @xmath2 of all galaxies have been corrected for evolution as derived from massive cluster galaxies , @xmath85 . at @xmath8 , the relation between @xmath2 and @xmath11 seems much steeper than in the local universe . however , we have to take selection effects into account . since we selected our sample in the @xmath18 band , we can transform our magnitude limit into a luminosity limit in the rest - frame @xmath54 band , which is close to the observed @xmath18 band at @xmath10 ( the median redshift of our sample of 16 @xmath10 galaxies is @xmath113 ) . because this luminosity limit only applies to our primary sample , this discussion does not involve the fillers ( mainly galaxies at @xmath114 ) . afterward , we will comment on these galaxies . we first test whether the observed distribution can be fully explained by selection effects , assuming that the slope and the scatter of the fp do not evolve . the probability that the observed distribution is drawn from a population with the same distribution as the sdss galaxies is 0.14% , if the @xmath2 evolution is the same for all galaxies . therefore , the slope evolves , the scatter evolves , or both . if there is an age difference between high and low mass galaxies , the scatter will most probably also evolve differently for high and low mass galaxies . we can not exclude with high confidence that the observed distribution of @xmath2 is due to a larger scatter at high redshift : if the scatter in @xmath2 at @xmath8 is twice as large as in the local universe , the probability of a non - evolving slope is 8.0% . however , we see no evidence for an increase in the scatter at the high mass end , where selection effects do not play a role . hence , if the scatter evolves , this is only true for galaxies with masses @xmath115 . since the increased scatter is most likely caused by young ages , low mass galaxies would have lower @xmath2 , and hence the slope of the relation would also be changed . we consider these findings as strong evidence for mass - dependent evolution of early - type galaxies . to confirm that we observed a change in the slope of the fp , deeper and larger surveys are needed . the location of the luminosity limit at @xmath8 in figure [ fig : m_ml ] shows that our @xmath10 sample is dominated by selection effects for masses @xmath116 . the objects which such low masses are only included in the sample because of their probably extreme @xmath2 . we can not correct the @xmath2 of that subsample for the bias introduced by our luminosity limit . on the other hand , for the sample of galaxies with higher masses , we can correct for selection effects , because they are relevant , but not dominant , as can be seen in figure [ fig : m_ml ] . the average evolution of the galaxies in the primary sample with masses @xmath117 is @xmath118 . the median mass of this subsample of 12 galaxies is @xmath119 . we estimate the maximum bias by assuming that the slope is the same at @xmath8 and in the local universe , but that the scatter is a factor two larger at @xmath8 than at @xmath120 . in that case the observed distribution is expected to follow the short - dashed line in figure [ fig : m_ml ] . at a given mass , the difference between the solid line and the short - dashed line is the bias introduced by the selection effects in luminosity . we increase the observed @xmath2 of each galaxy in our primary sample by the difference between the solid line and the short - dashed line at the mass of that galaxy . for galaxies more massive than @xmath121 this correction is negligible , but for the galaxies with masses @xmath122 this correction is about 30% . using this method , we find a bias corrected evolution of the galaxies with masses @xmath117 of @xmath123 . given the uncertainty in the intrinsic scatter , deeper observations are necessary to confirm this value . besides a bias due to the luminosity limit , errors in the velocity dispersion produce correlated errors in @xmath11 and @xmath2 , hence underestimating the measured evolution . taking the errors in @xmath53 into account , we find that this introduces a bias at the level of only 2 - 3% , which is several times smaller than our measurement accuracy . the above analysis only involves @xmath10 galaxies satisfying all our selection criteria , but we note that the relation between @xmath2 and @xmath11 exists for the @xmath114 galaxies as well . however , this subsample is selected in an inhomogeneous way , therefore it is impossible to correct the observed evolution . since all the galaxies roughly lie along lines of constant luminosity , the bias toward low @xmath2 galaxies likely explains the observed relation between @xmath11 and @xmath2 for this sub - sample . besides by luminosity and morphology , the galaxies in our sample are also selected by color . this potentially introduces an important bias in the measured evolution , because of the exclusion of galaxies with blue colors , i.e. , low @xmath2 . our color criterion , however , is quite generous . even the very blue , low mass galaxies satisfy this criterion . for typical bruzual - charlot models , our color limit ( @xmath124 ) corresponds to @xmath125 at @xmath8 , which is 0.2 mag bluer than the limit applied by bell et al . ( 2004b ) to select galaxies on the red sequence at @xmath10 . our color cut is 0.45 mag bluer than the color - magnitude relation found by the same authors . from bruzual - charlot models we estimate that we only miss galaxies that are younger than @xmath126 gyr ( see figures 7 and 8) . furthermore , bell et al . ( 2004b ) did not find blue , massive galaxies in the entire combo-17 dataset . hence it is very unlikely that we miss any blue galaxy at the bright end of our sample because of our color selection criterion . it is particularly interesting to compare our results to the results from studies involving lensing galaxies , since those samples are mass selected . rusin et al . ( 2003 ) and van de ven et al . ( 2003 ) find @xmath127 and @xmath128 , respectively , using the same dataset . this seems somewhat low compared to the evolution of our sample and results found in the literature , but the lensing galaxies typically have high masses ( see table 4 ) . the data for individual galaxies published by van de ven et al . show that the median mass of the lens sample is @xmath86 , whereas the median mass of our early - type galaxy sample is @xmath129 . the galaxies in the lens sample that are more massive than @xmath86 evolve as @xmath130 . the galaxies less massive than this evolve much faster : @xmath131 . the similarity between the results from the lensing sample and our sample is striking , especially because the lensing sample is mass selected . hence , it is very hard to see how a bias toward low @xmath2 galaxies can be responsible for the observed mass dependence in the lensing sample . it is possible that not all the low mass lenses are genuine early - types . if we omit the most irregular lenses ( fbq0951 + 2635 , sbs1520 + 530 , and b1608 + 656 ) , we still find rapid evolution , @xmath132 , for low mass lensing galaxies . hence , these results provide strong evidence that the observed dependence of @xmath2 evolution on mass in our sample is real . one possible complicating factor is that the lensing cross - section of galaxies in groups is larger than that of galaxies in the lowest density environments . this may lead to a difference between the populations of the lensing sample are our sample . it is interesting to see that galaxies with emission lines have relatively low @xmath2 , as can be seen from both figure [ fig : z_dml ] and [ fig : m_ml ] . treu et al . ( 2002 ) have already shown that a substantial fraction of the massive early - type galaxy population at high redshift shows evidence for ongoing star formation by the presence of emission lines . galaxies with emission lines in our sample , however , tend to have low masses . figure [ fig : m_ml ] suggests that several of the galaxies with masses @xmath133 are included in the magnitude limited sample only because they are forming stars . when excluding galaxies with emission lines , the evolution of the sample ( without applying a mass cut ) , is @xmath134 . the galaxies with emission lines evolve faster : @xmath135 . on average , the mass of the galaxies without emission lines is 1.5 times larger than the mass of the galaxies with emission lines . mergers or interactions , accompanied with star formation , can lead to both deviations from smooth @xmath49-profiles and low @xmath2 values . van dokkum & ellis ( 2003 ) have shown tentative examples of this phenomenon . figure [ fig : asym_dml ] shows the magnitude of the asymmetric residual , ( described in section [ sec : profiles ] ) , versus the offset from the fp , corrected for luminosity evolution . there is no correlation between deviations from smooth surface profiles and @xmath2 , and we find no evidence for a connection between star formation activity and interactions or mergers . four out of 11 early - type galaxies at @xmath136 have agn , and one out of 16 for the @xmath137 sample , as determined from x - ray imaging ( see section 3.3 ) . of the five early - type galaxies with agn , three are more massive than @xmath121 and two have emission lines in their spectra . the @xmath2 evolution of galaxies in our sample with agn is @xmath138 , which is not very different from the value for galaxies without agn : @xmath139 . however , the galaxies with agn are twice as massive than the galaxies without agn . considering high mass galaxies only ( @xmath95 ) , we find @xmath140 for galaxies with agn and @xmath141 for galaxies without agn . now , the trend is reversed , but only one out of the five galaxies with an agn has a particularly blue @xmath69 color for its @xmath2 . therefore , a larger sample is needed to confirm that the epochs of star formation and agn activity are related . woo et al . ( 2004 ) find that up to @xmath142 the evolution of the fp for galaxies selected to have agn is indistinguishable from otherwise selected galaxies . our results provide tentative evidence that a difference may set in at higher redshift . various other authors have measured the @xmath2 evolution of field early - type galaxies . in this section we compare these previous results with our findings , and we comment on the apparent inconsistencies that exist in the literature . in table 4 we give the @xmath2 evolution as reported by other papers , as well as the values we derive using the tabulated datasets given in those papers . treu et al . ( 2002 ) find @xmath143 . this number is corrected for selection effects , while the value in table 4 is uncorrected , to make a fair comparison with other results . the median mass of the sample presented by treu et al . ( 2003 ) is @xmath144 for a sample extending to @xmath145 . tabulated data are only available for the lower redshift galaxies presented in treu et al . only the galaxies with very low redshifts ( @xmath146 ) have low masses , therefore the comparison between high and low mass galaxies can not be made . we note that the evolution of the entire treu et al . sample is consistent with the evolution of our entire sample , @xmath147 , and that the evolution of the massive galaxies in our sample is only mildly inconsistent with the evolution of the treu et al several of the high mass galaxies in the treu et al . sample have emission lines in their spectra , which contrasts with our sample . van dokkum et al . ( 2001 ) find @xmath148 for an average redshift @xmath149 . the galaxy masses range from @xmath150 to @xmath151 and are evenly distributed in @xmath152 . remarkably , when taking the tabulated data of the individual galaxies , and applying our fitting method , we find a larger value : @xmath153 ( see table 4 ) . van dokkum et al . create redshift bins in which they calculate the bi - weight center of the @xmath2 offset . this method is sensitive to the bin choice , which leads to the difference between the reported value and the value calculated with our fitting method . we note that , when using our fitting method , the result of van dokkum et al . is very similar to the result presented in this paper . splitting the van dokkum et al . sample into low and high mass galaxies ( at @xmath86 ) , we find @xmath154 for the high mass galaxies , and @xmath155 for the low mass galaxies . we verified that this @xmath156 difference can be explained entirely by the fact that the galaxies are luminosity selected , not mass selected . contrary to this study and the work of treu et al . , van dokkum et al . did not select the galaxies by color , but by morphology and magnitude only . van dokkum & ellis ( 2003 ) find @xmath157 . the difference in fitting method , mentioned above , applies to this study as well . additionally , their transformation from @xmath110-band surface brightness to rest - frame @xmath54-band surface brightness , using @xmath158 colors , is uncertain . namely , using @xmath158 is an extrapolation for galaxies at @xmath159 . this leads to a different value for the @xmath2 evolution if we use our method to fit to the individual galaxy data of the samples of van dokkum et al . ( 2001 ) and van dokkum & ellis ( 2003 ) : @xmath160 , a value very similar to the result yielded by our sample . using the bi - weight center to measure the evolution effectively gives very low weighting factors to outliers . if the two galaxies with the lowest @xmath2 in the high-@xmath18 sample of van dokkum et al . are omitted , we find @xmath161 . the measurement accuracy , together with the uncertain transformation to rest - frame properties can explain the remaining difference . the new sample described by van dokkum & ellis ( 2003 ) is too small to verify whether there is a trend with galaxy mass . there is one galaxy more massive than @xmath86 with @xmath162 . as is the case for the sample of van dokkum et al . ( 2001 ) , the sample of van dokkum & ellis ( 2003 ) is not selected by color . this did not lead to faster evolution due to including blue early - type galaxies , which indicates that the color cuts in the other studies are adequately generous to avoid this bias . gebhardt et al . ( 2003 ) ( hereafter g03 ) report a brightening in the rest - frame @xmath54 band of early - type galaxies by 2.4 magnitudes at @xmath8 , which is derived from fitting a cubic spline to the offsets of the galaxies from the local fp . with our fitting method , we find @xmath163 , using the data of the individual galaxies in that sample . this is consistent with the result presented in this paper . the difference between the value reported by g03 and our result using their data is caused by the three galaxies at redshifts @xmath137 . our linear fit to their data and their fit agree well up to @xmath164 ( 1.6 magnitudes brightening ) . if we split the g03 sample into high and low mass galaxies , we find an evolution of @xmath165 for galaxies more massive than @xmath86 , and @xmath166 for less massive galaxies . for low mass galaxies our result , @xmath167 , and that of g03 agree well . the fast evolution for massive galaxies in their sample remains unexplained , but given the large uncertainty and the small number of objects the inconsistency is only mild . we note , however , that contrary to the early - type galaxies in our sample , the early - type galaxies in the g03 sample do not show a correlation between @xmath2 and rest - frame @xmath68 color . one of the galaxies in the g03 sample serves as an example of the uncertainty . + 5226 has been observed spectroscopically before ( ohyama et al . 2002 ) and happens to be in the sample of lensing galaxies of rusin et al . ( 2003 ) and van de ven ( 2003 ) . g03 give a stellar velocity dispersion of @xmath168 km s@xmath14 corrected to the same aperture as our dispersions . ohyama et al . have measured @xmath169 km s@xmath14 , also corrected to the same aperture , which implies a somewhat smaller offset from the local fp . van de ven et al . report a value of @xmath170 km s@xmath14 as derived from the lensing model . the latter value leads to a rate of evolution that we find for the massive galaxies in our sample . repeated observation of these massive galaxies may be illuminating . we conclude that all the results found in the literature are mutually consistent , once differences in calculating and presenting the results have been taken into account . we obtained ultra - deep spectroscopy for 27 field early - type galaxies with redshifts @xmath0 . the offset of these high redshift galaxies from the local fp is used as a measure of the evolution of @xmath2 and as an age estimator . the average evolution of the early type galaxies in our sample is @xmath90 . the value we find for galaxies in the primary sample , those galaxies satisfying all our selection criteria , is the same . the scatter in @xmath87 is large : 0.58 . this shows that some galaxies must have high luminosity weighted formation redshifts ( @xmath171 ) , while others have formed a large fraction of their stars at redshifts @xmath112 . emission lines in the spectra indicate that some galaxies show signs of ongoing star formation at the epoch of observation . this is in agreement with the presence of massive early - type galaxies with emission lines at @xmath142 treu et al . ( 2002 ) , although the galaxies with emission lines in our sample tend to have low masses . we find a tight correlation between @xmath2 and rest - frame color , which shows that the variation in @xmath2 among the galaxies in the sample is intrinsic , and due to differences in the stellar populations . the galaxies in our sample span a large range of masses . we find that low mass galaxies have larger offsets from the local fp than high mass galaxies . because luminosity selected samples are biased toward galaxies with low @xmath2 this is a trend that is expected . we carefully analyze whether the observed correlation between mass and @xmath2 can entirely be explained by this selection effect or not . we find that galaxies with masses @xmath172 are only included in our sample because they have low @xmath2 . for galaxies at @xmath10 with masses larger than @xmath173 , our sample is biased , but to a limited amount . taking into account the selection effect , we exclude with high confidence that the distribution of mass and @xmath2 of our @xmath10 galaxy sample with masses @xmath174 has the same distribution as the low redshift field early - type galaxy sample taken from bernardi et al . ( 2003 ) , corrected for evolution . we do not claim that we have observed a change of the slope of the fp , because we can not exclude the possibility that our sample is drawn from a distribution that has the same slope as the bernardi et al . sample , but with a scatter that is twice as large at @xmath8 . however , the outliers do not occur at the high mass end ( @xmath175 ) of our galaxy sample , but at lower masses , namely @xmath176 . therefore , our results show that the evolution of early - type galaxies is mass - dependent , whether by an increase in the scatter at lower masses , or by systematic faster evolution of lower mass galaxies as compared to higher mass galaxies , or , which is the most natural explanation , by a combination of these effects . assuming the scatter has decreased from @xmath8 to the present day by a factor of 2 , we find that the for bias corrected @xmath2 evolution of @xmath10 early - type galaxies with masses @xmath174 is @xmath123 . previous studies ( treu et al . 2001 , 2002 ; van dokkum et al . 2001 , van dokkum & ellis 2003 , gebhardt et al . 2003 ) , that claimed to have derived mutually exclusive results , are in fact consistent with our results , if the same fitting method is applied to the different datasets . particularly interesting is the consistency of our results with the results from a sample of lensing galaxies ( rusin et al . 2003 ; van de ven 2003 ) , which also shows the mass - dependence , even though this sample is not biased toward galaxies with low @xmath2 : because of the selection technique , the lens sample contains galaxies with typical @xmath2 at a given mass . our luminosity limited sample is sensitive to outliers , which are present indeed . the combination of these independent results strengthens the evidence for mass - dependent evolution and the combined increase in both slope and scatter with redshift . bell et al . ( 2004 ) claim that the mass density of red sequence galaxies increases by at least a factor of 2 from @xmath10 to the present . this is partly based on the observation that the luminosity density is constant out to @xmath8 . the @xmath2 evolution of our galaxy sample implies an increase by a factor of 4 in the mass density . in the local universe , most of the mass density in early - type galaxies is accounted for by galaxies with a velocity dispersion of @xmath177 km s@xmath14 ( kochanek et al . 2000 ) or a mass of @xmath178 . if those galaxies , which evolve somewhat slower , dominate the evolution of the mass density , the increase is slightly less ( @xmath179 ) . the correlation between mass and @xmath2 has been observed in clusters as well ( wuyts et al . 2004 ) , but this can entirely be explained by selection effects . both our field sample and the cluster samples found in the literature are not strongly biased for galaxy masses @xmath180 . when applying this mass cut , the evolution for cluster galaxies is @xmath181 , and the evolution for field galaxies is @xmath96 ( @xmath182 for galaxies in the primary sample ) . galaxies with masses comparable to the mass of @xmath183 galaxies in the local universe ( @xmath178 ) have luminosity weighted ages that imply formation redshifts @xmath111 , independent of environment . if progenitor bias is important , the luminosity weighted age of typical early - type galaxies in the local universe can be considerably lower . in hierarchical formation models , the predicted difference between the @xmath2 of field and cluster galaxies is @xmath184 , independent of redshift ( van dokkum et al . this large difference is related to the difficulty of constructing isolated galaxies without active star formation , i.e. to the lack of a mechanism that truncates star formation from within the galaxy . our results rule out this prediction at the 99.6% confidence level , up to @xmath185 . our findings are consistent with down - sizing ( cowie et al . 1996 ; kodama 2004 ) . this idea is independently corroborated by other observations , such as the decrease of the mass of e+a galaxies with time ( tran et al . 2003 ) , the lack of star formation in massive galaxies at redshifts @xmath186 ( de lucia et al . 2004 ) , the fossil record of star formation in local early - type galaxies ( thomas et al . 2004 ) , and the claim of mass - dependent evolution of spiral galaxies ( ziegler et al . 2002 , bhm et al . the lack of age differences between field and cluster galaxies , and the suggested mass - dependent evolution of early - type galaxies show that individual properties of a galaxy , and not environment , play an important role in its formation . we have shown that rest - frame optical colors can be used to measure galaxy masses at high redshift . this can be regarded as a step toward accurately calibrating sed fitting as a mass estimator . certainly including spitzer photometry in the rest - frame near - infrared will provide tight correlations between dynamically derived @xmath2 and @xmath2 derived from sed fitting . line - strengths of absorption features in our high @xmath34 spectra will connect low redshift fossil record studies to evolutionary studies such as these , and constrain the metallicity range of early - type galaxies at @xmath10 , lifting the age - metallicity degeneracy . using the mass calibration for high - redshift galaxies , the evolution of the mass density and the mass function can be determined from volume limited samples . this will provide strong constraints on formation theories and the importance of progenitor bias . we thank the referee for many useful comments , enhancing the quality and the readability of the work . we thank the eso staff for their professional and effective assistance during the observations . the lorentz center is thanked for its hospitality during various workshops and the leidsch kerkhoven - bosscha fonds for its financial support . alexander , d.m . , et al . , 2003 , , 126 , 539 bell , e. , mcintosh , d.h . , katz , n. , & weinberg , m.d . 2003 , , 149 , 289 bell , e.f . , et al . 2004 , , 608 , 752 bernardi . 2003 , , 125 . 1866 bertin , e. & arnouts , s. , 1996 , , 117 , 393 blakeslee , j.p . , et al . 2003 , , 596 , 143 bhm , a. et al . 2004 , , 420 , 97 bruzual , a. g. , charlot , s. 2003 , , 344 , 1000 cole , s. , lacey , s. , baugh , c.m . , frenk , c.s . 2000 , , 319 , 168 coleman , g.d . , wu , c .- c . , weedman , d.w . 1980 , , 43 , 393 cowie , l.l . , songaila , a. , hu , e.m . , & cohen , j.g . 1996 , , 112 , 839 de lucia , g. et al . 2004 , , 610 , l77 diaferio , a. , kauffmann , g. , balogh , m.l . , white , s.d.m . , schade , d. & ellingson , e. 2001 , , 323 , 999 djorgovski , s. & davis , m. 1987 , , 313 , 59 dressler , a. , lynden - bell , d. , burstein , d. , davies , r.l . , faber , s.m . , terlevich , r. , wegner , g. 1987 , , 313 , 42 faber , s.m . , wegner , g. , burstein , d. , davies , r. l. , dressler , a. , lynden - bell , d. , & terlevich , r. l. 1989 , , 69 , 763 kauffmann , g. , charlot , s. 1998 , , 279 , l23 franx , m , 1993 , , 105 , 1058 gebhardt , k. et al . 2003 , 597 , 239 giacconi , r. , et al . 2002 , 139 , 369 giavalisco . 2004 , , 600 , l93 holden , b.p . , et al . 2005 , , in press , astro - ph/0412570 jrgensen , i. , franx , m. & kjrgaard , p. 1995 , , 276 , 134 jrgensen , i. , franx , m. & kjrgaard , p. 1996 , , 280 , 167 kauffmann , g. , & charlot , s. 1998 , , 297 , l23 kelson , d.d . , illingworth , g.d . , franx , m. , van dokkum , p.g . 2000 , , 531 , 184 kochanek , c.s . 2000 , , 543 , 131 kodama et al . 2004 , , 350 , 1005 krist , j. 1995 , in asp conf . 77 , astronomical data analysis software and systems iv , ed . shaw , h.e . payne , & j.j.e . hayes ( san fransisco : asp ) , 349 lidman , c. , rosati , p. , demarco , r. , nonino , m. , mainieri , v. , stanford , s.a . , toft , s. 2004 , , 416 , 829 ohyama , y et al . 2002 , , 123 , 2903 rosati , p. , et al . 2004 , , 127 , 230 rusin , d. et al . 2003 , , 587 , 143 thomas , d. , maraston , c. , bender , r. , & c. mendez de oliveira 2004 , , in press tran , k .- v . , franx , m. , illingworth , g.d . , kelson , d.d . , & van dokkum , p.g . 2003 , , 599 , 865 treu , t. , stiavelli , m. , bertin . g. , casertano , s. , & mller , p. 2001 , , 326 , 237 treu , t. , stiavelli , m. , casertano , s. , mller , p. & bertin , g. 2002 , , 564 , l13 valdes , f. , gupta , r. , rose , j.a . , singh , h.p . , bell , d.j . 2004 , , 152 , 251 vandame , b. et al . , in preparation van der wel . a. , franx , m. , van dokkum , p.g . , rix , h .- w . , 2004 , , 601 , l5 van de ven , p.m. , van dokkum , p.g . , franx , m. 2003 , , 344 , 924 van dokkum , p.g . 2001 , , 113 , 1420 van dokkum , p.g . , franx , m. 1996 , , 281 , 985 van dokkum , p.g . , franx , m. , kelson , d.d . , illingworth , g.d . 1998 , , 504 , l17 van dokkum , p.g . , franx , m. , kelson , d.d . , illingworth , g.d . 2001 , , 553 , l39 van dokkum , p.g . , franx , m. 2001 , , 553 , 90 van dokkum , p.g . , ellis , r.s . 2003 , , 592 , l53 van dokkum , p.g . & stanford , s.a . 2003 , s.a . 2003 , , 585 , 78 wolf , c. et al . 2004 , , 421 , 913 woo , j.h , urry , c.m . , lira , p. , van der marel , r.p . , and maza , j. 2004 , , in press wuyts , s. , van dokkum , p.g . , kelson , d.d . , franx , m. , illingworth , g.d . 2004 , , 605 , 677 ziegler , b.l . 2002 , , 564 , 69 lccccccccccccc i d & e & x - ray & @xmath187 & @xmath34 & @xmath53 & type & @xmath51 & @xmath188 & @xmath21 & @xmath25 & @xmath189 & @xmath67 + & & & & & km s@xmath14 & & @xmath190 & & & & & + cl1252 - 1 & & a & 0.671 & 48 & @xmath191 & k2 & 0.46 & 21.69 & 20.72 & 0.53 & 17.27 & 1.95 + cl1252 - 2 & & & 0.658 & 75 & @xmath192 & g2 & 0.18 & 20.04 & 20.35 & 0.46 & & + cl1252 - 3@xmath193 & @xmath194 , [ ] & a & 0.844 & 83 & @xmath195 & f6 & 1.04 & 22.78 & 19.30 & 0.54 & & + cl1252 - 4@xmath193 & [ ] & & 0.743 & 93 & @xmath196 & f8 & 1.02 & 23.00 & 19.56 & 0.41 & 17.49 & 1.58 + cl1252 - 5 & & a & 0.743 & 61 & @xmath197 & k1 & 0.36 & 21.07 & 19.91 & 0.53 & 16.98 & 1.81 + cl1252 - 6 & & & 0.734 & 123 & @xmath198 & g2 & 0.24 & 20.43 & 20.19 & 0.51 & 17.51 & 1.77 + cl1252 - 7 & [ ] & a & 0.753 & 135 & @xmath199 & g8 & 0.33 & 20.53 & 19.56 & 0.47 & & + cl1252 - 8 & [ ] & & 1.069 & 19 & @xmath200 & f4 & 0.27 & 21.92 & 21.41 & 1.00 & & + cl1252 - 9 & & & 1.036 & 18 & @xmath201 & f8 & 0.22 & 21.71 & 21.59 & 0.88 & & + cdfs-1 & [ ] & a & 1.089 & 27 & @xmath202 & f4 & 0.44 & 21.97 & 20.36 & 1.12 & 17.70 & 1.83 + cdfs-2 & & & 0.964 & 40 & @xmath203 & g0 & 0.39 & 21.47 & 20.10 & 1.04 & 17.15 & 1.74 + cdfs-3 & & & 1.044 & 17 & @xmath204 & g4 & 0.15 & 20.39 & 21.14 & 1.07 & 18.29 & 1.75 + cdfs-4 & & x & 0.964 & 33 & @xmath205 & g0 & 1.00 & 22.94 & 19.55 & 1.11 & 16.44 & 1.89 + cdfs-5 & & & 0.685 & 32 & @xmath206 & k1 & 0.37 & 21.64 & 20.42 & 0.66 & 18.27 & 1.81 + cdfs-6 & & & 0.660 & 49 & @xmath207 & g0 & 0.15 & 19.36 & 20.10 & 0.59 & 17.32 & 1.67 + cdfs-7 & & & 1.135 & 17 & @xmath208 & g4 & 0.74 & 23.32 & 20.58 & 1.14 & 17.26 & 1.97 + cdfs-8 & & & 1.125 & 6 & @xmath209 & g4 & 0.20 & 21.53 & 21.66 & 1.19 & 18.70 & 1.84 + cdfs-9 & & & 1.097 & 8 & @xmath210 & g0 & 0.21 & 21.20 & 21.21 & 1.22 & 18.09 & 1.86 + cdfs-10 & [ ] & & 1.119 & 9 & @xmath211 & f6 & 0.091 & 19.68 & 21.50 & 1.10 & 18.72 & 1.70 + cdfs-11 & [ ] & & 1.096 & 10 & @xmath212 & g2 & 0.23 & 21.87 & 21.67 & 1.23 & 17.93 & 1.9 + cdfs-12 & & & 1.123 & 20 & @xmath213 & g4 & 0.14 & 20.42 & 21.29 & 1.17 & 18.00 & 1.87 + cdfs-13 & & & 0.980 & 46 & @xmath214 & g4 & 0.20 & 20.93 & 20.17 & 1.05 & 17.22 & 1.83 + cdfs-14 & & & 0.984 & 18 & @xmath215 & g4 & 0.39 & 22.19 & 20.85 & 1.01 & 17.25 & 1.78 + cdfs-15 & & & 0.622 & 55 & @xmath216 & g8 & 0.31 & 20.94 & 20.09 & 0.62 & 17.41 & 1.73 + cdfs-16 & & & 0.669 & 31 & @xmath217 & g8 & 0.33 & 21.09 & 20.12 & 0.66 & 17.32 & 1.81 + cdfs-17@xmath193 & & & 0.954 & 18 & @xmath218 & f4 & 0.83 & 23.33 & 20.35 & 0.89 & 17.84 & 1.81 + cdfs-18 & & & 1.096 & 14 & @xmath219 & g0 & 0.51 & 22.21 & 20.3 & 1.20 & 16.98 & 1.92 + cdfs-19 & [ ] & & 0.955 & 15 & @xmath220 & g4 & 0.34 & 20.35 & 19.30 & 0.49 & 17.53 & 1.56 + cdfs-20 & & & 1.022 & 25 & @xmath221 & g2 & 0.34 & 21.69 & 20.65 & 1.19 & 16.94 & 1.93 + cdfs-21 & & & 0.735 & 56 & @xmath222 & f4 & 0.073 & 18.52 & 20.81 & 0.56 & 18.05 & 1.59 + cdfs-22 & & a , x & 0.735 & 46 & @xmath223 & k1 & 0.67 & 22.25 & 19.73 & 0.76 & 16.84 & 1.77 + cdfs-23 & & & 1.041 & 13 & @xmath224 & g4 & 0.39 & 22.59 & 21.23 & 1.09 & 18.51 & 1.77 + cdfs-24@xmath193 & & a & 1.042 & 25 & @xmath225 & f8 & 1.87 & 24.35 & 19.60 & 0.96 & 17.10 & 2.00 + cdfs-25 & & & 0.967 & 28 & @xmath226 & f6 & 0.13 & 20.02 & 21.10 & 1.02 & 18.13 & 1.65 + cdfs-26@xmath193 & & & 1.129 & 14 & @xmath227 & g0 & 1.12 & 23.83 & 20.19 & 1.14 & 17.43 & 1.99 + cdfs-27 & [ ] & & 1.128 & 8 & @xmath228 & f6 & 0.67 & 23.68 & 21.15 & 0.91 & 18.69 & 1.58 + cdfs-28@xmath193 & [ ] & a & 0.954 & 12 & @xmath229 & g4 & 0.80 & 23.54 & 20.64 & 0.80 & 18.52 & 1.78 + cdfs-29 & & & 1.128 & 19 & @xmath230 & f6 & 0.21 & 20.89 & 20.86 & 1.12 & 17.58 & 1.65 + lcccccccc i d & @xmath187 & @xmath231 & @xmath232 & @xmath152 & @xmath233 & @xmath234 & @xmath68 & @xmath69 + & & @xmath235 & @xmath236 & @xmath237 & @xmath238 & & & + cl1252 - 1 & @xmath239 & @xmath240 & @xmath241 & @xmath242 & @xmath243 & @xmath244 & @xmath245 & @xmath246 + cl1252 - 2 & @xmath247 & @xmath248 & @xmath249 & @xmath250 & @xmath251 & @xmath252 & @xmath253 & + cl1252 - 3@xmath193 & @xmath254 & @xmath255 & @xmath256 & @xmath257 & @xmath258 & @xmath259 & @xmath260 & + cl1252 - 4@xmath193 & @xmath261 & @xmath262 & @xmath263 & @xmath264 & @xmath265 & @xmath266 & @xmath267 & @xmath268 + cl1252 - 5 & @xmath261 & @xmath269 & @xmath270 & @xmath271 & @xmath109 & @xmath272 & @xmath245 & @xmath273 + cl1252 - 6 & @xmath274 & @xmath275 & @xmath276 & @xmath277 & @xmath278 & @xmath279 & @xmath280 & @xmath281 + cl1252 - 7 & @xmath282 & @xmath283 & @xmath284 & @xmath285 & @xmath286 & @xmath287 & @xmath288 & + cl1252 - 8 & @xmath289 & @xmath290 & @xmath291 & @xmath292 & @xmath293 & @xmath294 & @xmath295 & + cl1252 - 9 & @xmath296 & @xmath297 & @xmath298 & @xmath299 & @xmath300 & @xmath301 & @xmath302 & + cdfs-1 & @xmath303 & @xmath304 & @xmath305 & @xmath306 & @xmath307 & @xmath308 & @xmath309 & @xmath310 + cdfs-2 & @xmath311 & @xmath312 & @xmath313 & @xmath314 & @xmath286 & @xmath315 & @xmath251 & @xmath316 + cdfs-3 & @xmath317 & @xmath318 & @xmath319 & @xmath320 & @xmath321 & @xmath322 & @xmath323 & @xmath324 + cdfs-4 & @xmath311 & @xmath325 & @xmath326 & @xmath327 & @xmath328 & @xmath329 & @xmath330 & @xmath331 + cdfs-5 & @xmath332 & @xmath333 & @xmath334 & @xmath335 & @xmath336 & @xmath337 & @xmath94 & @xmath331 + cdfs-6 & @xmath338 & @xmath339 & @xmath340 & @xmath341 & @xmath258 & @xmath342 & @xmath278 & @xmath343 + cdfs-7 & @xmath344 & @xmath345 & @xmath346 & @xmath347 & @xmath348 & @xmath349 & @xmath309 & @xmath350 + cdfs-8 & @xmath351 & @xmath352 & @xmath353 & @xmath354 & @xmath355 & @xmath356 & @xmath357 & @xmath310 + cdfs-9 & @xmath358 & @xmath359 & @xmath360 & @xmath361 & @xmath362 & @xmath363 & @xmath364 & @xmath365 + cdfs-10 & @xmath366 & @xmath367 & @xmath368 & @xmath369 & @xmath370 & @xmath371 & @xmath307 & @xmath372 + cdfs-11 & @xmath373 & @xmath374 & @xmath375 & @xmath376 & @xmath321 & @xmath377 & @xmath330 & @xmath378 + cdfs-12 & @xmath379 & @xmath380 & @xmath381 & @xmath382 & @xmath362 & @xmath383 & @xmath384 & @xmath385 + cdfs-13 & @xmath386 & @xmath387 & @xmath388 & @xmath389 & @xmath258 & @xmath390 & @xmath391 & @xmath316 + cdfs-14 & @xmath392 & @xmath393 & @xmath394 & @xmath395 & @xmath364 & @xmath396 & @xmath309 & @xmath310 + cdfs-15 & @xmath397 & @xmath398 & @xmath399 & @xmath400 & @xmath401 & @xmath402 & @xmath403 & @xmath404 + cdfs-16 & @xmath405 & @xmath406 & @xmath407 & @xmath408 & @xmath328 & @xmath409 & @xmath251 & @xmath410 + cdfs-17@xmath193 & @xmath411 & @xmath412 & @xmath413 & @xmath414 & @xmath415 & @xmath416 & @xmath258 & @xmath385 + cdfs-18 & @xmath373 & @xmath417 & @xmath418 & @xmath419 & @xmath420 & @xmath421 & @xmath422 & @xmath423 + cdfs-19 & @xmath424 & @xmath425 & @xmath426 & @xmath427 & @xmath428 & @xmath429 & @xmath430 & @xmath431 + cdfs-20 & @xmath432 & @xmath433 & @xmath434 & @xmath435 & @xmath258 & @xmath436 & @xmath403 & @xmath437 + cdfs-21 & @xmath438 & @xmath439 & @xmath440 & @xmath441 & @xmath430 & @xmath442 & @xmath278 & @xmath443 + cdfs-22 & @xmath438 & @xmath444 & @xmath445 & @xmath446 & @xmath447 & @xmath448 & @xmath321 & @xmath449 + cdfs-23 & @xmath450 & @xmath451 & @xmath452 & @xmath453 & @xmath454 & @xmath455 & @xmath357 & @xmath456 + cdfs-24@xmath193 & @xmath457 & @xmath458 & @xmath459 & @xmath460 & @xmath420 & @xmath461 & @xmath286 & @xmath462 + cdfs-25 & @xmath463 & @xmath464 & @xmath465 & @xmath466 & @xmath295 & @xmath467 & @xmath468 & @xmath469 + cdfs-26@xmath193 & @xmath470 & @xmath471 & @xmath472 & @xmath473 & @xmath267 & @xmath474 & @xmath307 & @xmath475 + cdfs-27 & @xmath476 & @xmath477 & @xmath478 & @xmath479 & @xmath480 & @xmath481 & @xmath482 & @xmath483 + cdfs-28@xmath193 & @xmath411 & @xmath484 & @xmath485 & @xmath486 & @xmath487 & @xmath488 & @xmath489 & @xmath490 + cdfs-29 & @xmath476 & @xmath491 & @xmath492 & @xmath493 & @xmath494 & @xmath495 & @xmath480 & @xmath496 + lccccccc ref . & @xmath497 & @xmath497 & @xmath497 & @xmath497 & @xmath498 & @xmath499 & @xmath500 + & reported & fitted & high mass & low mass & & & + treu et al . 2001 & @xmath501 & @xmath502 & @xmath503 & @xmath504 & @xmath505 & 0.29 & 19 + van dokkum et al . 2001 & @xmath506 & @xmath507 & @xmath508 & @xmath509 & @xmath510 & 0.42 & 18 + treu et al . 2001 , 2002 & @xmath511 & & & & @xmath512 & 0.38 & 29 + van dokkum et al . 2001 , 2003 & @xmath513 & @xmath514 & @xmath515 & @xmath516 & @xmath517 & 0.56 & 27 + gebhardt et al . 2003 & @xmath518 & @xmath519 & @xmath520 & @xmath521 & @xmath522 & 0.64 & 21 + this paper & @xmath523 & @xmath523 & @xmath524 & @xmath525 & @xmath526 & 0.90 & 27 + rusin et al . 2003 & @xmath527 & & & & & 0.54 & 21 + van de ven et al . 2003 & @xmath528 & @xmath529 & @xmath530 & @xmath531 & @xmath505 & 0.54 & 21 +	we present an analysis of the fundamental plane for a sample of 27 field early - type galaxies in the redshift range @xmath0 in the chandra deep field - south and the field of the background cluster rdcs 1252.9 - 2927 . sixteen of the galaxies are at @xmath1 . the galaxies in this sample have high signal - to - noise spectra obtained at the very large telescope and high resolution imaging from the hst advanced camera for surveys . from comparison with lower redshift data , we find that the mean evolution of the mass - to - light ratio ( @xmath2 ) of our sample is @xmath3 , with a large galaxy - to - galaxy scatter . this value can be too low by 0.3 due to selection effects , resulting in @xmath4 . the strong correlation between @xmath2 and rest - frame color indicates that the observed scatter is not due to measurement errors , but due to intrinsic differences between the stellar populations of the galaxies , such that our results can be used as a calibration for converting luminosities of high redshift galaxies into masses . this pace of evolution is much faster than the evolution of cluster galaxies . however , we find that the measured @xmath2 evolution strongly depends on galaxy mass . for galaxies with masses @xmath5 , we find no significant difference between the evolution of field and cluster galaxies : @xmath6 for field galaxies and @xmath7 for cluster galaxies . the relation between the measured @xmath2 evolution and mass is partially due to selection effects , as the galaxies are selected by luminosity , not mass . however , even when taking selection effects into account , we still find a relation between @xmath2 evolution and mass , which is most likely caused by a lower mean age and a larger intrinsic scatter for low mass galaxies . results from lensing early - type galaxies , which are mass - selected , show a very similar trend with mass . this , combined with our findings , provides evidence for down - sizing , i.e. , for the proposition that low mass galaxies are younger than high mass galaxies . previous studies of the rate of evolution of field early - type galaxies found a large range of mutually exclusive values . we show that these differences are largely caused by the differences between fitting methods : most literature studies are consistent with our result and with one another when using the same method . finally , five of the early - type galaxies in our sample have agn . there is tentative evidence that the stellar populations in these galaxies are younger than those of galaxies without agn .
electric vehicles ( evs ) are emerging as a sustainable and environmentally friendly alternative to conventional vehicles , provided that the energy used for their charging is obtained from renewable energy sources . the energy generated from renewable sources such as sunlight , wind and waves is , however , dependent on weather conditions . as a consequence , the electricity production from these sources is inherently uncertain in time and quantity . furthermore , electricity has to be produced and consumed at the same time , as the large - scale storage of the energy generated is , still today , very limited . as a result , the energy obtained from renewables may be wasted in times when the demand for electricity is not high enough to absorb it , with a consequent detrimental effect on the profitability of renewables . since the battery in an ev is basically a storage device for energy , the large - scale integration of evs in the transportation sector may contribute to substantially increasing the socioeconomic value of an energy system with a large renewable component , while reducing the dependence of the transportation sector on liquid fossil fuel . for this reason , evs have received increased interest from the scientific community in recent years ( detailed literature reviews of the state of the art can be found in @xcite and @xcite ) . special attention has been given to the analysis of the effect of evs integration on the electricity demand profile @xcite , emissions @xcite and social welfare @xcite , and to the design of charging schemes that avoid increasing the peak consumption @xcite , help mitigate voltage fluctuations and overload of network components in distribution grids @xcite , and/or get the maximum economic benefit from the storage capability of evs within a market environment , either from the perspective of a single vehicle @xcite or the viewpoint of an aggregator of evs @xcite . in all these publications , though , and more generally in the technical literature on the topic , the charging problem of an ev is addressed either by considering deterministic driving patterns , when the focus is placed on the management of a single vehicle , or by aggregating the driving needs of different ev users , when the emphasis is on modeling a whole fleet of evs . this aggregation , however , obscures the dynamics of each specific vehicle . likewise , the deterministic driving patterns of a single ev are often based on expected values or stylized behaviors , which fail to capture important features of the charging problem such as the daily variation in the use of the vehicle or potential user conflicts in terms of not having the vehicle charged and ready for use . a stochastic model for driving patterns provides more insight into these aspects and becomes fundamental for applying a charging scheme in the real world . despite this , the stochastic modeling of driving patterns has received little attention from the scientific community , as pointed out in @xcite . we mention here the research work by @xcite , in which they aim to capture the uncertainty intrinsic to the vehicle use by means of a monte carlo simulation approach . they assume , however , an uncontrolled charging scheme . the work developed in this paper departs from the following two premises : 1 . the primary purpose of the battery of an ev is to provide power to drive the vehicle and not to store energy from the electricity grid . consequently , it is essential that enough energy is kept in the battery to cover any desired trip . this calls for a decision tool that takes into account the driving needs of the ev user to determine when charging can be postponed and when the battery should be charged right away . 2 . the complexity of human behavior points to a stochastic model for describing the use of the vehicle . in turn , this stochastic model should be integrated into the aforementioned decision tool and exploited by it . that being so , this paper introduces an algorithm to optimally decide when to charge an ev that exhibits a stochastic driving pattern . the algorithm builds on the inhomogeneous markov model proposed in @xcite for describing the stochastic use of a single vehicle . the model parameters are then estimated on the basis of data from the use of the specific vehicle . the approach captures the diurnal variation of the driving pattern and does not rely on any assumptions on the use of the vehicle , which makes it general and particularly versatile . our algorithm thus embodies a _ markov decision process _ which is solved recursively using a stochastic dynamic programming approach . the resulting decision - support tool allows for addressing issues related to charging , vehicle - to - grid ( v2 g ) schemes @xcite , availability and costs of using the vehicle . the algorithm runs swiftly on a personal computer , which makes it feasible to implement on an actual ev . the remainder of this paper is organized as follows : in section 2 the stochastic model for driving patterns developed in @xcite is briefly described , tailored to be used in the present work , and extended to address the problem of driving data limitations through hidden markov models . section 3 introduces the algorithm for the optimal charging of an ev as a markov decision process that is solved using stochastic dynamic programming . section 4 provides results from a realistic case study and explores the potential benefit of implementing v2 g schemes . section 5 concludes and provides directions for future research within this topic . in this section we summarize and extend the stochastic model for driving patterns developed in @xcite . we refer the interested reader to this work for a detailed description of the modeling approach . a state - space model is considered to describe the use of the ev . in its simplest form , it contains two states , according to which the vehicle is either _ driving _ or _ not driving_. a more extensive version of the model would include a larger number of states which could capture information about where the vehicle is parked , how fast it is driving or what type of trip it is on . the basics of the general multi - state stochastic model are described in this section , including how to fit a specific model on an observed data set . let @xmath0 , where @xmath1 , be a sequence of random variables that takes on values in the countable set @xmath2 , called the state space . denote this sequence as @xmath3 . we assume a finite number , @xmath4 , of states in the state space . a markov chain is a random process where future states , conditioned on the present state , do not depend on the past states @xcite . in discrete time @xmath3 is a markov chain if @xmath5 for all @xmath6 and all @xmath7 . a markov chain is uniquely characterized by the transition probabilities , @xmath8 , i.e. @xmath9 if the transition probabilities do not depend on @xmath10 , the process is called a homogeneous markov chain . if the transition probabilities depend on @xmath10 , the process is known as an inhomogeneous markov chain . when it comes to the use of a vehicle , it is appropriate to assume that the probability of a transition from state @xmath11 to state @xmath12 is similar on specific days of the week . thus , for instance , thursdays in different weeks will have the same transition probabilities . for convenience we further assume that all weekdays ( monday through friday ) have the same transition probabilities . these assumptions can be easily relaxed or interchanged with other assumptions and as such , are not essential to the model . with a sampling time in minutes , and taking into account that there are 1440 minutes in a day , this leads to the assumption : @xmath13 this assumption implies that the transition probabilities , defined by ( [ transprob ] ) , are constrained to be a function of the time , @xmath14 , in the diurnal cycle . let the matrix containing the transition probabilities be denoted by @xmath15 . for the model containing @xmath4 states the transition probability matrix is given by : @xmath16 where @xmath17 . now let @xmath18 define the number of observed transitions from state @xmath11 to state @xmath12 at time @xmath14 . from the conditional likelihood function , the maximum - likelihood estimate of @xmath19 can then be found as : @xmath20 a discrete time markov model can be formulated based on the estimates of @xmath21 . one apparent disadvantage of such a discrete time model is its huge number of parameters , namely @xmath22 , where @xmath23 parameters have to be estimated for each time step . needless to say , the number of parameters to be estimated increases as the number of states grows . we refer to @xcite for further details on techniques to reduce the number of parameters to be estimated for each time step for models with more than two states . another problem is linked to the number of observations available to properly carry out the estimation , i.e. if @xmath24 for some @xmath25 , then @xmath26 is undefined . to deal with the large number of parameters as well as undefined transition probability estimates , b - splines are applied to capture the diurnal variation in the driving pattern through a _ generalized linear model_. the procedure of applying a _ generalized linear model _ is implemented in the statistical software package r as the function ` glm(\cdot ) ` . for a thorough introduction to b - splines see @xcite and for a general treatment of generalized linear models see @xcite . next we elaborate on how the fitting of the markov chain model works in our particular case . each day , at a specific minute , a transition from state @xmath11 to state @xmath12 either occurs or does not occur . thus for every @xmath14 on the diurnal cycle we can consider the number of transitions to be binomially distributed , i.e. @xmath27 , where the number of bernoulli trials at @xmath14 , given by @xmath28 , is known and the probability of success , @xmath19 , is unknown . the data can now be analyzed using a _ logistic regression _ , which is a generalized linear model @xcite . the explanatory variables in this model are taken to be the basis functions for the b - spline . the logit transformation of the odds of the unknown binomial probabilities are modeled as linear combinations of the basis functions . we model @xmath29 and in particular , we are interested in @xmath30 = p_{jk}(s)$ ] . as the basis functions for the b - spline are uniquely determined by the knot vector @xmath31 , deciding the knot position and the amount of knots is important to obtain a good fit for the model . here we proceed as follows : first a number of knots are placed on the interval @xmath32 $ ] , with one at each endpoint and equal spacing between them . denote this initial vector of knots by @xmath33 . the model is then fitted using the basis functions as explanatory variables . next , the fit of the model between the knots is evaluated via the likelihood function and an additional knot is placed in the center of the interval with the lowest likelihood value . the new knot vector is then given by @xmath34 . we repeat this procedure until the desired number of knots is reached . to determine the appropriate number of knots and avoid over - parametrization , on the basis of a likelihood ratio principle , we test that adding a new knot does significantly improve the fit . standard markov models are limited in the sense that only those states that are actually observed can be modeled . thus , if the data at our disposal only provide information on when the vehicle is either _ driving _ or _ not driving _ , the standard markov model is restricted to two states . furthermore the time spent in each state is exponentially distributed , albeit with time - varying intensity , and accordingly , the time until the next transition does not depend on the time spent in the current state . this may be particularly unrealistic for a model with few states for describing driving patterns . to overcome these limitations , we can use a hidden markov model , which allows estimation of additional states that are not directly observed in the data . in fact , we can estimate these states so that the waiting time in each state matches that which is actually observed in the data . adding a hidden state is done by introducing a new state in the underlying markov chain . the new state , however , is indistinguishable from any of the previously observed states . this allows for the waiting time in each observable state to be the sum of exponential variables , which is a more versatile class of distributions . it is worth insisting that the use of hidden markov models is justified here to address insufficient state information in our data , which only include whether the vehicle is _ driving _ or _ not driving_. indeed , the same results could be obtained using the underlying markov chain without hidden states , provided that the hidden states could be observed . in practice , though , more detailed driving data ( e.g. including driving speed and/or location of the vehicle ) could be available once the actual implementation is made on a vehicle , which in turn would avert the need for a hidden markov model . for a detailed introduction to hidden markov models , see @xcite , where techniques and scripts for estimating parameters are also provided . the hidden markov model consists of two parts . firstly , an underlying unobserved markov process , @xmath35 , which describes the actual state of the vehicle . this part corresponds to the markov model with no hidden states as described previously . the second part of the model is a state - dependent process , @xmath36 , such that when @xmath0 is known , the distribution of @xmath37 depends only on the current state @xmath0 . a hidden markov model is thus defined by the state - dependent transition probabilities , @xmath8 , as defined for the standard markov chain and the state - dependent distributions given by ( in the discrete case ) : @xmath38 collecting the @xmath39 s in the matrix @xmath40 , the likelihood of the hidden markov model is given by : @xmath41 where @xmath42 is the initial distribution of @xmath43 . we can now maximize the likelihood of observations to find the estimates of the transition probabilities . the data at our disposal is from the utilization of a single vehicle in denmark in the period spanning the six months from 23 - 10 - 2002 to 24 - 04 - 2003 , with a total of 183 days . the data is gps - based and follows specific cars . one car has been chosen and the model is intended to describe the use of this vehicle accordingly . the data set only contains information on whether the vehicle was _ driving _ or _ not driving _ at any given time . no other information was provided in order to protect the privacy of the vehicle owner . the data is divided into two periods , a training period for fitting the model from 23 - 10 - 2002 to 23 - 01 - 2003 , and a test period from 24 - 01 - 2003 to 24 - 04 - 2003 for evaluating the performance of the model . the data set consists of a total of 749 trips . the time resolution is in minutes . we shall consider a model with one _ not driving _ state and several ( hidden ) _ driving _ states . that is , it is observable if the vehicle is driving , but not the specific driving state . to fit the model to the data , we assume that only the transition probability from the _ not driving _ state depends on the time of day . this is done to reduce the complexity of the estimation procedure , as it is cumbersome to estimate the time - varying parameters of a hidden markov model . it is worth noting that a hidden markov model allows for the probability of ending the current trip to depend on the time since departure , as the vehicle may pass through different driving states before ending the trip . we now elaborate on the fitting of the hidden markov model , which is split into estimation of its time - varying and time - invariant parameters . we need to estimate the probability of a transition from the vehicle being parked to a driving state . we denote this transition estimate by @xmath44 . it holds that @xmath45 . since both the parked state and the transitions from it are directly observable in the data , we can use the procedure described in section 2.1 to estimate @xmath44 . + the data have been divided into two main periods : weekdays and weekends . the observed number of trips starting every minute for the weekdays is displayed in fig . [ startingweekdays ] . a high degree of diurnal variation is found , with a lot of trips starting around 06:00 and again around 16:00 . also , there are no observations of trips starting between 00:00 and 05:00 . other patterns are found for weekends , but as these do not involve any methodological difference , we limit ourselves to trips starting on weekdays . annual variations may also be present , however the limited data sample does not allow for capturing such seasonality . based on the b - splines and the logistic regression , plotted as the black line over the estimates @xmath44 from ( [ p_hat ] ) , in gray . the red bars indicate the knot positioning.__,scaledwidth=80.0% ] the plot in fig . [ transitionprobability ] illustrates the estimate of @xmath44 using b - splines with eight initial knots placed uniformly on the interval and 22 knots in total . the time - invariant parameters are to be estimated so that an appropriate probability distribution is fitted to the duration of the trips . the time - invariant parameters are estimated by maximizing the likelihood given in ( [ eq : hmm likelihood ] ) . for a given number of _ driving _ states , the transition probabilities can be estimated using the approach in @xcite . once a model with @xmath4 states is fitted , we can test if adding an additional state significantly improves the fit . as a model with @xmath4 states is a sub - model of one with @xmath46 or more states , we increase the number of states until no significant improvement test is observed according to the likelihood ratio . [ triplength ] represents the histogram of the empirically observed trip lengths along with the theoretical density function of the trip lengths obtained from the fitted model . we use a model with two driving states , as no significant improvement is found beyond this number . notice that the distribution of the empirically observed trip lengths is adequately captured by the hidden markov model , although the number of observed trips in the range from 10 to 20 minutes has a higher prevalence than the fitted distribution . in practical applications , more information could be available to model the behavior of the vehicle ( e.g. its location and speed ) , which should facilitate the modeling of the driving patterns . in the following section , the algorithm for optimally charging the ev is presented . the optimization algorithm makes use of the transition probabilities characterizing the stochastic model for the driving patterns . thus , the optimization algorithm is designed to handle the stochastic nature of the driving needs . the problem of charging an ev can be posed as a conflict between two opposing objectives . the end - user desires to have the vehicle charged and ready for use at his / her discretion , while also minimizing the costs of running the vehicle . demand for electricity varies over the day and so does the electricity generated from renewable sources . this introduces a varying energy price which can make it beneficial for the end - user to postpone charging his / her vehicle . this means the user is faced with the problem of postponing charging to minimize costs or to charge right away so as to maximize the availability of the vehicle . the algorithm for optimal charging of the ev is formulated as a stochastic dynamic programming problem . we first define the relevant parameters and variables , and then the state - transition and objective function . [ cols= " < " , ] let us consider first the strategies in tab . [ tab : outofsample ] under which only charging is allowed , we see that there are no observed events of not having enough charge on the battery to complete a trip . also , we notice , as expected , that the optimal charging strategies have lower costs than the `` rule of thumb '' policies . the low - price charging strategy is indeed the `` rule of thumb '' policy that approximates closest to the optimal policy in terms of costs and availability . it yields , however , an average daily cost which is around 12 - 24% higher than that obtained from implementing the proposed decision - support tool . as for the charging policies that include v2 g operation mode , it becomes apparent that caution should be exercised to prevent the vehicle from being fully discharged when the end - user desires to drive . in this line , notice that increasing the penalty reduces the number of observed events of not having enough charge on the battery to cover a desired trip . introducing a v2 g charging scheme allows for substantially reducing the cost associated with driving as opposed to charging - only schemes , and may even result in negative average costs . observe that the optimal charging policy developed in this paper clearly outperforms the `` rule of thumb '' v2 g schemes . in the unbounded case , charging costs are substantially reduced , but multiple out - of - battery events are recorded . imposing a lower bound on the discharging solves this problem , but at the expense of considerably increasing the running cost of the vehicle , to such an extent that it nearly doubles . the difference in performance between the optimal charging strategy and the `` rule of thumb '' policies can be expected to become larger for electric vehicles covering higher distances or with lower battery capacity . lastly , we would like conclude this section by pointing out that , in general , the spot price is not the price observed by the end - user . indeed , the end - user faces a price that includes taxes and other costs on top of the spot electricity price . as an example , consider a country like denmark , where the average price of electricity paid by the end - user , including taxes and fees , is around 300 /mwh , which is 5 - 10 times the average spot price @xcite . in the current danish power system , fees and taxes are imposed on the amount of electricity consumed by the end - user , not on its total cost . this does not encourage the end - user to switch to a smart consumption of energy based on variable prices . in fact , if the taxes and fees were implemented as a function of the total energy cost , the savings from switching to a smart charging policy in denmark could be multiplied by a factor of between five and ten . this paper proposes an algorithm to optimally charge an electric vehicle based on stochastic dynamic programming . the algorithm is built on an inhomogeneous ( hidden ) markov chain model that characterizes the stochastic use of the vehicle . the algorithm determines the optimal charging policy depending on the use of the vehicle , the risk aversion of the end - user , and the electricity price . the costs associated with running the vehicle are decreased significantly when the charging strategy is determined by the proposed optimization model , with little or no inconvenience to the end - user . these costs can be reduced even further if the vehicle is permitted to supply power into the grid . indeed , findings show the possibility of making a net profit from running the vehicle . the proposed stochastic dynamic programming model is versatile and can easily be adapted to any specific vehicle , thus providing a customized charging policy . a possible extension would be to apply the proposed model to data with more markov states , which could be used to investigate the benefits of installing more public charging stations as opposed to home charging , or to capture different driving states such as `` urban '' , `` rural '' , or `` highway '' . in addition , the model could be enhanced to consider transition probabilities that are estimated adaptively in time . an adaptive approach would capture structural changes in the driving behavior , such as variations over the year or a change in use that could follow , for example , from the householder buying an additional vehicle . moreover , adaptivity is relevant for applying the model in practice . further research could be also directed at modeling a fleet of vehicles by using a mixed - effects model . the optimization scheme could be applied individually to each vehicle and the total population load could be evaluated . this would highlight if and how evs could be used to mitigate an increase in peak electricity demand when switching from combustion - based vehicles to evs . other investigations could focus on the relationship between evs and renewable energy sources and how evs could be used to move the excess production to time periods of high demand , possibly making renewables more economically competitive . dsf ( det strategiske forskningsrd ) is to be acknowledged for partly funding the work of emil b. iversen , juan m. morales and henrik madsen through the ensymora project ( no . 10 - 093904/dsf ) . furthermore , juan m. morales and henrik madsen are partly funded by the ipower platform project , supported by dsf ( det strategiske forskningsrd ) and rti ( rdet for teknologi og innovation ) , which are hereby acknowledged . finally , we thank dtu transport for providing the data used in this research . 23 natexlab#1#1[1]`#1 ` [ 2]#2 [ 1]#1 [ 1]http://dx.doi.org/#1 [ ] [ 1]pmid:#1 [ ] [ 2]#2 , , , , in : , , pp . . , , , ( ) , , , , ( ) . , , ( ) . , , , in : , , pp . . , , , in : , , , pp . . , , , ( ) . , , , , , california institute of technology , . . , , , , in : , , pp . . , , , , ( ) . , , , ( ) . , , , ( ) . , , , , ( ) . , , , , in : , , pp . . , , , , , in : , , pp . . , , , , , , technical university of denmark , . . , , , ( ) . , , , texts from oxford university press , , . , , , , springer series in statistics , , . , , , chapman & hall / crc texts in statistical science , , . , , , monographs on statistics and applied probability , , .	the combination of electric vehicles ( evs ) and renewable energy is taking shape as a potential driver for a future free of fossil fuels . however , the efficient management of the ev fleet is not exempt from challenges . it calls for the involvement of all actors directly or indirectly related to the energy and transportation sectors , ranging from governments , automakers and transmission system operators , to the ultimate beneficiary of the change : the end - user . an ev is primarily to be used to satisfy driving needs , and accordingly charging policies must be designed primarily for this purpose . the charging models presented in the technical literature , however , overlook the stochastic nature of driving patterns . here we introduce an efficient stochastic dynamic programming model to optimally charge an ev while accounting for the uncertainty inherent to its use . with this aim in mind , driving patterns are described by an inhomogeneous markov model that is fitted using data collected from the utilization of an ev . we show that the randomness intrinsic to driving needs has a substantial impact on the charging strategy to be implemented . electric vehicles , driving patterns , optimal charging , markov processes , stochastic dynamic programming ,
systems biology has rapidly advanced to offer a number of powerful methods for discovery of networks of genes / proteins and collective functions of families of genes / proteins in life processes . analysis of quantitative traits loci ( qtl ) refers to a systematic method for discovery of genes and their functions in systems biology . the major bottleneck in qtl progress appears to be in the more general and challenging question of how to quantify phenotypic traits , such as morphology and dynamics of its variation . qtl and other gene function discovery methods rely on quantification of phenotypic traits that could be observed and used in a systematic way to distinguish between organisms with differing dna sequences or epigenetic signatures . to extract phenotypic traits that distinguish genotypic characteristics , biologists make repeated observations of the wild type and the mutants of the same species during growth , behavior or in the course of response to external stimuli . in the following , we illustrate and outline the steps for identifying and quantifying phenotypic traits of seedling of the model plant _ arabidopsis thaliana _ subject to changes in gravitational force relative to its natural root orientation during a normal course of growth . it will be demonstrated that the selected quantitative traits are together sufficiently informative to carry genomic signatures that are at work differently in the wild type and mutant plants . plant roots vary a great deal in morphology , size and complexity of their architecture . the model plant _ arabidopsis thaliana _ has a root system that is exemplary in both prevalence of its morphological pattern and its relative simplicity . there are several visible morphological features of this root system . the primary root is the main branch that starts its life immediately after germination . besides being a significant method for simplifying morphological measurement , midline extraction serves as an example of high - dimensional data transformation that achieves dimensionality reduction with some loss of information beyond noise , namely , ignoring the signal - parameters underlying phenotypic variation such as morphology of the root due to image texture or image structures such as shadows of root hair . in higher resolution images , the interpretation will change , and root hairs could also be part of the features whose morphological diversity and variation in growth dynamics become essential biological quantities to be studied that are related to specific gene functions , pathways or gene - protein network dynamics . in this article we provide details of a set of algorithms for extraction of root and hair growth information . tropism refers to the directed growth responses of plants to external stimuli such as gravity , water , light , and temperature . beginning with the pioneering work of darwin , the study of tropism has grown to a major area of research in plant biology in the course of its century - long history . darwin observed that grass seedlings grown in dark tend toward a light source when illuminated from one side . plant roots show a similar response to gravity . when a seed germinates , the roots penetrate the soil and grow downward . however , if a root is reoriented by 90@xmath0 with respect to the gravitational field , the root responds by altering its direction of growth , curving until it is again vertical . sachs was the first one to propose a quantitative measure for gravitropism @xcite , namely , the gravitropic response was proportional to the component of the gravity vector perpendicular to the root axis . early studies , using maize roots , demonstrated distinct regions along the root axis with different physiological response patterns . despite this long and illustrious research history , the molecular mechanisms involved in sensing the gravitational signal and its transduction still need to be studied . other tropic responses are sources of gaining information about the physiological and molecular processes that influence the tropic response , e.g. the plant hormone auxin has been implicated in tropic responses through extensive research by a broad group of scientists ( @xcite ) . auxin is involved in asymmetric tropic growth , vascular development and root formation , as well as in a plethora of other processes . the approaches by which auxin has been implicated in tropisms include isolation of mutants altered in auxin transport or response with altered gravitropic or phototropic response , identification of auxin gradients with radiolabeled auxin and auxin - inducible gene reporter systems , and by use of inhibitors of auxin transport that block gravitropism and phototropism . proteins that transport auxin have been identified and the mechanisms which determine auxin transport polarity have been explored . the mechanisms of auxin action in the gravitropic response and phototropism have recently been revealed by the analysis of mutants that are defective in the response to the gravity signal , or have a characteristic response to blue and other spectral bands of light . gravitational stimulus induces curvature of the primary root of _ arabidopsis thaliana_. it is known that the patterns of root curvature reveal quantitative traits that are associated to a number of genes and proteins that play vital roles in growth and development of the plant . the application described below gives an outline of machine - learning methods that distinguish mutant from wild type seedling plants in gravitropism experiments . we note that these applications could be adapted for other experimental protocols in plant biology that attempt to quantify subtle phenotypic traits in order to decipher functions of genes - proteins . we have designed and engineered a complete system , specifically focusing on high throughput plant imaging . this hardware - software product is a novel _ `` portable modular system for automated image acquisition '' _ in the lab and for certain field experiments . our model system is described in @xcite . the algorithms below are part of `` the image analyzer '' component of our system . it is an object - oriented software application that can be modified to accommodate essentially any reasonable analysis scenario . the image analyzer package is developed in linux , and it is ready for experimental protocols by the biologist , see @xcite . this platform can provide high throughput results in a range of resolutions , and accommodates highly flexible experimental protocols . this article demonstrates the feasibility of using our system towards fully automated high throughput phenotyping in functional genomics and systems biology ( figure [ fig1 ] ) . parts of a high resolution image of growth of lateral roots and extraction of morphological through the algorithms developed by the authors . [ d ] the midlines of primary root images frames ( please see below ) could be metaphorically positioned on top of each other to visualize the dynamics of growth in gravitropism as the geometry of the surface . we have developed algorithms for extraction of midlines of root hair and dynamics of growth as the next step in capturing phenotypic variation . ] in earlier contributions by the senior author and liya wang ( @xcite , @xcite ) image analysis and midline extraction were only partially automated and required _ `` preprocessing tasks '' _ prior to application of the software . preprocessing steps are highly data - dependent image processing steps . in this article , we report successful automation of the preprocessing , and a novel collection of algorithms that are also amenable to generalization beyond the plant roots , as well as parallelization . the preprocessing methods are adapted from earlier seminal research by osher et al ( @xcite , called the total variation regularization ( tv ) method . we remark that comparable approaches have been also treated in literature using bounded variation ( bv ) . as expected , direct application of one set of programs to another typically shows varying degrees of success . therefore , it is anticipated that generalization of the automation preprocessing would depend on the setting for image acquisition and would vary in degrees of successful analysis in plant growth dynamics of various kinds . the hardware and automation offered by cyplant solutions inc . have a standardized setting that removes a series of technical issues in order to ensure the quality of data for the same genre of mathematical treatment as in our algorithms and software . our work in progress will provide a critical examination of a number of algorithms and methods in literature , and in appropriate cases , redesigns the needed image preprocessing algorithms from scratch to be optimal for the tasks , such as those in cyplant solutions standardized technologies . we next present the outline of the algorithm . let the image for one gray - scale frame be represented by @xmath1 the basic idea is as follows : the typical image @xmath2 is written as an outcome of an unknown convolution kernel @xmath3 with compact support ( e.g. a discrete gaussian ) , applied to the original image @xmath4 ( or a desirable model that in our case must be suitable for segmentation and midline recovery of the branched structures , cf . @xcite ) , and an additive noise @xmath5 , that we propose to model as a gaussian white noise , such that @xmath6 the automation algorithm selects the kernel and a suitable model for noise ( that we must confront in practical applications ) , which will be derived from the high throughput imaging system outputs in lieu of the simplifying assumption above . we will begin with the constrained minimization problem with appropriate norms , without further mention of simplicity . for a class of images obtained by our method , the higher resolution allowed us to select a delta function , and use the simpler form . however , in the case of animal behavior , blurring occurs due to movement of the animal or temporal loss of the automated focus . in the case of plants , blurring occurs due to accumulation of moisture on inside surfaces of the petri dish covers . this means that we must use the general form below : @xmath7 here , @xmath8 is a scaling constant that is the trade - off between noise and the desired image quality , such as having sharp edges in images of root systems , and arg - min is over all images @xmath9 , where @xmath4 is the desired form of the image suitable for algorithms of @xcite ( for an iterative regularization to recover finer scales ) . an intermediate problem that we solved numerically was the parallel implementation of methods based on the following idea that is adapted from a simplified mathematical result by chan - wong - kaveh - osher et al . that amounts to a constrained optimization problem with @xmath10 as a function of joint variables @xmath11 : @xmath12 when we view @xmath10 as a one - variable functional depending only on @xmath3 or only @xmath9 , by fixing the other one , then @xmath10 is convex , while @xmath13 fails to be jointly convex . to overcome this difficulty , we followed @xcite and instead numerically solved the two euler - lagrange equations , where the @xmath14-conjugates are denoted by the `` hat '' : @xmath15 fixing @xmath3 first , we solve for @xmath9 from the first equation and then switch the roles of @xmath3 and @xmath9 and exchange the two equations . for a class of images according to our protocol , the higher resolution allows one to select an iterative scheme that mimics convergence to the appropriate delta function , which in turn allows one to use the simpler form mentioned above . however , in the case of movement of the entire root by sliding on the agar surface , blurring sometimes occurs ; this could also be a result of temporal loss of the automated focus . as mentioned before , in the case of plants , blurring mostly occurs due to accumulation of moisture on inside surfaces of the petri dish covers . in our companion article @xcite , and in our 2008 worldcomp article @xcite , and subsequent developments @xcite , @xcite , @xcite , we have developed a set of new algorithms and their object - oriented c - code for massively parallel - distributed hardware . this part of research was done in collaboration with nvidia and sun network.com . having quantified morphological features from the data set of movies , we continue to discover candidates to serve as phenotypic traits that carry the plant genotypic signature that account for distinguishing between the wild type and the mutants . a number of these features depend on computation of the midline of root images as argued in @xcite . the algorithm for midline tracing in the present work is different from the main algorithm in @xcite . in fact , the second generation of midline tracing algorithms used in the present article is more robust , more accurate and much faster @xcite . we computed midlines from different regions of a root image as follows . gravitropism results in anisotropic expansion of cell walls and an uneven change in the distribution of cells in the epidermis of the root in such a way that the root growth tends to generate the observed bending with respect to its growth in the original vertical direction @xcite . the part of root before the bending region is referred to as the _ horizontal region_. the bending region itself is called the _ hook region _ and finally , the part of the root after the hook region is called the _ vertical region_. lengths of these regions are calculated using the midline in these regions , and it serves as a feature . besides the length associated to the above - mentioned regions , the curve made by the midline carries the information about the root curvature during its gravitropic growth . further considerations ( omitted here ) indicate that the number of segments account for the changes in the growth direction which could potentially be a significant representative feature . the greatest change in the growth direction in the hook region is called the _ hook angle_. in addition to the geometric features of the midline of the primary root growth , we also developed algorithms to extract growth information regarding the root hairs . discussion of the biological significance and the algorithms for root hair growth information are available in @xcite . let us only mention that the number of root hairs for each region of the root is also a feature that we added to the list of morphological features for the classification purposes . using the root hair information , we formulate additional morphological features which we call _ dynamic features _ because they correspond to the growth velocity and growth acceleration of the primary root and the root hairs . figure [ fig2 ] shows these extracted features . all of the following features highlighted in the figure are considered as representative features of the growth images : * length of the vertical region , * length of the horizontal region , * length of the hook region , * the number of segments , * the hook angle , * the average root hair length , * the number of root hairs , * the root growth velocity , * root growth acceleration , * hair growth velocity , * hair growth acceleration , and * hair density . from a human observer s standpoint , the slow rate of growth of the root morphological features results in `` invisible phenotypic traits '' that belong to the complex dynamical system underlying the development and growth . in @xcite we have developed the preliminary steps towards the in - depth study of plant development within the framework of complex dynamical systems . for comprehensive details of these algorithms readers are referred to visit our website for @xcite and @xcite . to recap , we have used image analysis algorithms and mathematical modeling of dynamics of morphological features in order to isolate a number of quantitative morphological and dynamical features , and using these features , every movie from the growth process is assigned a representation by the appropriate vector , as described earlier . the final step towards discovery of quantitative phenotypic traits requires machine learning for classifying the above - mentioned features , and to train the machine through a set of so - called ` training samples ' or ` examples ' in order to extract the appropriate weighted combination of features that could carry genotypic signatures of the mutation ( here , _ mdr1 _ , which is one of the members of the _ mdr _ gene family , is knocked to provide the mutant ) . the machine learning method of choice here comes from the branch of statistical learning theory , called support vector machine ( svm ) , see @xcite . the flexibility of the svm theory has proved advantageous here , because we tailored to our needs an rbf - based svm , where rbf refers to the `` radial basis functions '' as our choice for the svm kernel function . we applied the rbf - svm method to 281 seedling growth movies ( 161 mutant _ mdr1 _ and 120 wild type seedlings ) where 12 features captured them as described before . we used the matlab - svm classifier @xcite and calculated the precision of our classification outcome . table [ tab1 ] shows the results for each class of movies . .this table shows the result of applying the svm method on the data set , where the radial basis function was used as the kernel function . precision of the results shows efficiency of using the defined features for capturing different classes of genotypic attitudes . [ cols="<,^,^ " , ] the precision was computed via the standard error formula @xmath16 since twelve features were employed we could not show a full @xmath17-dimensional figure of the classes . instead , the projection of the movies on the two coordinates corresponding to the features `` number of segments '' and `` average hair growth acceleration '' is shown in figure [ fig3 ] . among higher plants , the model organism _ arabidopsis thaliana _ has been studied in detail . as its genome is sequenced and better understood than other similar plants , there are many opportunities for investigation of the genotype - phenotype mapping . the development of the arabidopsis root system is a centerpiece in plant biology . due to the wide availability of a large amount of literature on root development , the arabidopsis root system continues to serve as an excellent model organ to investigate systems biology of higher plants . the functional landscape of gene - protein network dynamics is believed to be responsible for the regulation of growth and development of roots . multi - scale mathematical modeling of root growth provides a critical element for a systematic study of the mechanisms of regulatory transcription networks that operate on different scales , specific interactions of numerous proteins and tightly intertwined protein interaction networks . gene expression studies need to be performed at temporal and spatial resolutions ( minutes and micrometers ) that are relevant to the dynamics of gene and phenotype crosstalk . historically , tropic growth responses have been at the center of such activities . tropic responses allow plants to redirect their growth in response to their surrounding environment . the temporal dynamics of root growth is relatively accurately measured by following the displacement of features at extremities , such as the tips of roots and root hairs . it is necessary to obtain velocity profiles of neighboring elements as they are moved by expansion @xcite . velocity profiles can be obtained by imaging a growing organ over time which provides measures of the position of the externally applied marks , thus computing velocity as a function of position @xcite . the mathematical analysis of image sequences goes back to numerous early investigators , e.g. @xcite , and algorithmic methods substitute for laborious and often subjective manual measurements ( @xcite ) . several improvements were made by developing algorithms that measure the spatial growth profiles using image - processing techniques that could also utilize cell borders , intercellular air spaces or other physiological structures that are followed throughout a sequence of images ( see @xcite ) . concepts originally developed by @xcite and @xcite and improved by @xcite establish quantitative relationships for curvature production and curvature angle distribution . nevertheless , there is a continuing need to develop better algorithms that work more accurately in automated high throughput imaging systems such as our system outlined above . high throughput plant imaging systems require appropriate software applications for accurate automated image analysis . we have developed a prototype hardware - software product that operates as a _ `` portable modular system for automated image acquisition and analysis '' _ in the lab and some field applications . this article provides evidence for feasibility of using our system towards fully automated high throughput phenotyping in challenging functional genomics and systems biology applications . the methods described above also point to the many opportunities that machine learning could offer plant functional biology , such as distinguishing mutant from wild type seedling plants in gravitropism experiments . as we observed , these applications could be adapted for other experimental protocols in plant biology that attempt to quantify subtle phenotypic traits in order to decipher the functions of genes - proteins . the general problem of quantifying _ all _ phenotypic traits ( for example , in tropism ) for the development and growth of the plant root system remains a formidable challenge that is certain to inspire new insights in machine learning and analysis of massive biological imaging data . s. balasubramanian , c. schwartz , a. singh , n. warthmann , m. c. kim , j. n. maloof , o. loudet , g. t. trainer , t. dabi , j. o. borevitz , j. chory , and d. weigel . qtl mapping in new arabidopsis thaliana advanced intercross - recombinant inbred lines . , 4(2):e4318 , 02 2009 . h. dashti , a. ardalan , and a. assadi . a complex dynamical system approach to modeling of the arabidopsis root gravitropism with applications to discovery and quantification of invisible phenotypic traits . preprint . h. t. dashti , m. e. kloc , t. simas , r. a. ribeiro , and a. h. assadi . , volume 314 of _ ifip advances in information and communication technology _ , chapter introduction of empirical topology in construction of relationship networks of informative objects , pages 3543 . springer boston , 2010 . j. n. maloof , j. o. borevitz , t. dabi , j. lutes , r. b. nehring , j. l. redfern , g. t. trainer , j. m. wilson , t. asami , c. c. berry , d. weigel , and j. chory . natural variation in light sensitivity of arabidopsis . , 29:441446 , 2001 . j. mullen , e. turk , k. johnson , c. wolverton , h. ishikawa , c. simmons , d. sll , and m. l. evans . root - growth behavior of the arabidopsis mutant rgr1 : roles of gravitropism and circumnutation in the waving / coiling phenomenon . , 118(4):11391145 , 1998 . s. savaldi - goldstein , t. baiga , f. pojer , t. dabi , c. butterfield , g. parry , a. santner , n. dharmasiri , y. tao , m. estelle , j. noel , and j. chory . new auxin analogs with growth - promoting effects in intact plants reveal a chemical strategy to improve hormone delivery . , 105(39):1519015195 , 2008 . d. schmundt , m. stitt , b. jhne , and u. schurr . quantitative analysis of the local rates of growth of dicot leaves at a high temporal and spatial resolution , using image sequence analysis . , 16(4):505514 , 1998 . j. tonejc , h. torabi , and a. assadi . mathematical modeling of morphological dynamics in development and growth of the arabidopsis and maize seedlings with applications to quantifying phenotypic variation in their root system . , 2010 . a. walter , h. spies , s. terjung , r. kusters , n. kirchgessner , and u. schurr . spatio - temporal dynamics of expansion growth in roots : automatic quantification of diurnal course and temperature response by digital image sequence processing . , 53(369):689698 , 2002 . l. wang , i. v. uilecan , a. h. assadi , c. a. kozmik , and e. p. spalding . : image analysis software for measuring hypocotyl growth and shape demonstrated on arabidopsis seedlings undergoing photomorphogenesis . , 149(4):16321637 , 2009 .	post - genomic research deals with challenging problems in screening genomes of organisms for particular functions or potential for being the targets of genetic engineering for desirable biological features . ` phenotyping ' of wild type and mutants is a time - consuming and costly effort by many individuals . this article is a preliminary progress report in research on large - scale automation of phenotyping steps ( imaging , informatics and data analysis ) needed to study plant gene - proteins networks that influence growth and development of plants . our results undermine the significance of phenotypic traits that are implicit in patterns of dynamics in plant root response to sudden changes of its environmental conditions , such as sudden re - orientation of the root tip against the gravity vector . including dynamic features besides the common morphological ones has paid off in design of robust and accurate machine learning methods to automate a typical phenotyping scenario , i.e. to distinguish the wild type from the mutants . * keywords : * gravitropism , machine learning , phenotypic traits
charge ordered structure of @xmath9-(edt - ttf - conme@xmath10)@xmath10x . cations are omitted for clarity . high charge edt - ttf - conme@xmath10 molecules are drawn by red , low charge molecules by blue . antiferromagnetic heisenberg chains are parallel with _ @xmath11 is the strong antiferromagnetic coupling between spins along ( _ a _ ) , @xmath12 is the small anisotropic exchange interaction . [ fig : co_structure],width=321 ] the interest in quantum spin chains arises to a large extent from the exact theoretical solutions available for one - dimensional systems . furthermore , the unusual phenomena related to quantum fluctuations@xcite are more significant in one dimension than in higher dimensions . among the various experimental methods to study the spin dynamics of strongly correlated quantum systems , electron spin resonance ( esr ) is particularly useful as it is sensitive to anisotropic interactions . earlier esr studies of one - dimensional magnetic systems focused mainly on short range fluctuations at low temperatures and at low frequencies @xcite . however , appropriate theoretical models were lacking and the activity faded away in the mid-1970s . as a result of recent progress in many - body theory , exact solutions are now available , which predict the properties of the esr spectra in @xmath2 quantum antiferromagnetic chains @xcite . the calculations are for chains where a staggered field alternating along the chain is added to a weakly anisotropic antiferromagnetic exchange interaction that is uniform for all molecules . in real materials there is always some exchange anisotropy or a dipolar field . staggered effective fields may arise for structural reasons , e.g. the g - factor tensor alternates in a chain of molecules with alternating orientation . more interestingly , in chains of identical yet dissymmetrical molecules , not related by an inversion symmetry , a uniform external field , @xmath13 induces an effective staggered field due to the dzyaloshinskii - moriya ( dm ) interaction . calculations yield well defined predictions for the esr spectrum . the exchange anisotropy contributes a temperature dependent term to the esr line width , @xmath8 . a staggered field with an amplitude proportional to @xmath13 contributes terms proportional to @xmath14 to the g - factor and @xmath15 . in contrast to the wealth of new theoretical results , experimental studies are lagging behind . detailed multi - frequency esr measurements below 200 ghz frequency exist for quasi-1d inorganic chain compounds @xcite and copper pyrimidine dinitrate@xcite . here we present esr measurements of the quasi - one - dimensional ( q1d ) organic compounds ( x = asf@xmath0,br ) up to 420 ghz . these are promising systems for a study of s=1/2 quantum magnet chains . they are based on a dissymmetrical , acentric organic @xmath16-donor molecule ( fig . [ fig : co_structure ] ) . the radical cations stacked along _ a _ form the antiferromagnetic chain . the molecules are uniformly spaced along the chain and the strictly quarter filled hole band@xcite foreshadows a metallic conductivity . surprisingly , the compounds are insulating@xcite ; strong nearest neighbor repulsive coulomb interactions lead to a charge ordered ( co ) state@xcite . in the co pattern ( fig . [ fig : co_structure ] ) , recently identified by high resolution x - ray diffraction @xcite in the br compound , high charge and nearly neutral low charge molecules alternate in the _ a _ and _ b _ directions . a charge ratio of 0.9__e _ _ : 0.1__e _ _ was determined by nmr@xcite . the @xmath17 space group unit cell at room temperature is non - centrosymmetric . both systems undergo a continuous orthorhombic to monoclinic phase transition below @xmath18 k. at low temperatures they are three dimensionally ordered antiferromagnets , a peak in the nmr spin lattice relaxation rate at 8 and 12 k in the asf@xmath19 and br compounds respectively were attributed to the nel temperature by auban - senzier et al . a detailed low temperature esr study is in progress . here we present multi - frequency esr measurements in the high temperature orthorhombic @xmath17 phase of ( x = asf@xmath0,br ) in order to avoid complications due to structural and magnetic phase transitions . we suggest that the charge - rich molecules are coupled through charge - poor molecules to form antiferromagnetic chains along _ a_. we estimate a small uniform exchange anisotropy @xmath20 . the unusual quadratic frequency dependence of the esr line width and resonance field is attributed to a staggered effective field from a weak dzyaloshinskii moriya antisymmetric exchange interaction permitted by the non - centrosymmetric crystal structure . all experiments were on needle - like single crystals , grown electro - chemically as reported earlier@xcite . the longest edge of the crystal is parallel to the direction of radical cation stacks ( _ a _ ) , and the shortest crystal direction is along the longest unit cell axis ( _ c _ ) . for esr studies in the millimeter - wave region , the crystals were oriented under an optical microscope with a precision of approximately @xmath21 using this morphology . in the low - frequency @xmath22ghz ) esr measurements we used a goniometer and determined the crystallographic orientation from the extremal g - value and @xmath8 orientations . the precision is about @xmath23 with this later method . the low - frequency @xmath24 , @xmath25 and @xmath26 ghz esr measurements were done on commercial bruker elexsys super - s , e500 , and super - q spectrometers , respectively . the maximum magnetic field for these measurements was @xmath27 t. the temperature was controlled in the @xmath24 and @xmath26 ghz esr measurements with an oxford instruments cf935p cryostat and at @xmath25 ghz with an oxford esr900 cryostat . measurements in the millimeter - wave frequency range were carried out on a home - made quasi - optical continuous wave esr spectrometer . the available microwave frequencies were @xmath28 , @xmath29 and @xmath30 ghz . since this setup uses broadband microwave components only , it allows for frequency dependent esr measurements in a fixed probe head by simply changing the microwave radiation frequency . the magnetic field up to @xmath31 t was provided by an oxford instruments superconducting solenoid . we controlled the sample temperature in the @xmath32 k range in a continuous - flow cryostat . for precise g - value determination , we measured the microwave frequency with a hp-5352b frequency counter and the magnetic field was calibrated with polymeric kc@xmath33 powder ( @xmath34 ) . more details about the spectrometer are given in ref . there is a single , symmetric lorentzian line in all directions and at all frequencies above the antiferromagnetic ordering temperatures . we measured @xmath8 and the g - shift relative to the free electron g - value @xmath35 at fixed frequencies as a function of temperature , and also performed frequency dependent measurements at fixed temperature in order to clarify the nature of the small perturbation @xmath36 which may arise from some exchange anisotropy or from staggered effective fields . at low frequencies @xmath8 and @xmath35 were determined by fitting the measured angular dependence to a cosine function . at high - frequencies we performed temperature and frequency dependent measurements with field fixed along the main axes . data were fitted to the following parabolic expansions to describe phenomenologically the frequency and temperature dependence : @xmath37 where , @xmath38 are the three crystallographic orientations . first , in section [ ss : susc ] we discuss the temperature dependence of the spin susceptibility in the high temperature orthorhombic phase . then we summarize the theoretical predictions ( section [ ss : theory ] ) on the effect of the exchange anisotropy and the staggered field on the esr line width @xmath39 and g - shift of an @xmath40 antiferromagnetic heisenberg - chain . finally , in sections [ ss : low - freq ] and [ ss : high - freq ] we present the low- and high - frequency esr behavior respectively of @xmath9-(edt - ttf - conme@xmath10)@xmath10x ( x = asf@xmath0 , br ) systems . we restrict the discussion to temperatures above 190 k. below this temperature the continuous change of the structure ( and thus the exchange parameters ) over about 100 k complicates the analyses @xcite . ghz for the asf@xmath19 and br crystals , open red and closed green circles , respectively . lines are the bonner - fisher calculation with @xmath3 k and @xmath4 k. [ fig : asf6:br : susc],width=321 ] the spin - susceptibility at @xmath25 ghz is plotted as a function of temperature in fig . [ fig : asf6:br : susc ] . a fit to the bonner - fisher curve @xcite shows that in the high temperature orthorhombic phase above @xmath18 k ( the onset temperature of a second order orthorhombic to monoclinic phase transition @xcite ) the susceptibility follows the theoretical prediction for @xmath2 antiferromagnetic heisenberg chains . the coupling constants determined from the fits are @xmath41 k and @xmath4 k. the model describes well the temperature dependence and also the magnitude of the measured susceptibility for both salts . thus antiferromagnetic heisenberg chains are the characteristic magnetic building blocks of the systems . the chains run most probably along the _ a _ axis , overlap in other directions is much smaller . the first neighbor overlap integrals were calculated@xcite for uniformly spaced chains . the ratio of the integrals in the _ a _ and _ b _ directions is about 3 in both br and asf@xmath19 compounds . there is no sizeable overlap in the third ( _ c _ ) direction . in the co state , the anisotropy of the electron structure is not known ; the q1d character of the spin susceptibility suggests that it is large . the exchange interactions , @xmath42 , determined from the bonner - fisher calculations are surprisingly strong taking into account that essentially only every second molecule carries a spin . before discussing the esr spectrum further , we summarize theoretical predictions for a heisenberg antiferromagnetic chain with a small anisotropy . the hamiltonian of the chain is @xmath43 where @xmath36 is a small anisotropic perturbing interaction , @xmath44 runs along the _ a _ axis . @xmath13 is the external magnetic field . in the absence of anisotropic terms , the exchange , @xmath45 , does not shift the esr from the paramagnetic resonance frequency @xmath46 and does not broadens the line either . the perturbation @xmath36 due to an exchange anisotropy or a staggered effective field shifts and broadens the esr line . the temperature dependence of the esr line width of an anisotropic exchange coupled chain with @xmath47 , where an anisotropy axis is along _ p _ , was calculated@xcite in the @xmath48 parameter range : @xmath49 where @xmath50 for an applied field @xmath51 perpendicular ( parallel ) to the anisotropy axis , @xmath52 . the behavior is t - linear at low temperatures . we included only the largest temperature dependent correction to the linear dependence . at the limit of the validity of the field theory calculation@xcite , @xmath53 , there is a smooth crossover to the high - temperature kubo - tomita limit of a constant line width of the order of @xmath54 . if the perturbation is a staggered magnetic field with amplitude @xmath55 alternating along the chain , @xmath56 , then the resulting broadening and shift@xcite are : @xmath57 the effective staggered field may arise from an alternating g - tensor or the dm interaction . in these cases the perturbation amplitude , @xmath58 is proportional to the external field and @xmath59 depends on the direction of applied field . the field dependence distinguishes the broadening and shift due to @xmath60 from the field independent effects of the uniform exchange or dipolar anisotropy of @xmath61 . here again , we included only the largest term of the low temperature approximation . the high temperature limit of the line width arising from a staggered field is not well known . oshikawa and affleck@xcite proposed that it is qualitatively different for an alternating g - factor and dm interactions . in the former case the well known motional narrowing argument applies and the limiting line width is of the order of @xmath62 . on the other hand , for the dm interaction there is a field independent and field dependent term , proportional to @xmath63 and @xmath64respectively . and g - shift @xmath65 of measured at @xmath25 ghz in the @xmath66 and @xmath67 planes . @xmath68 is the angle between the _ b _ axis and field @xmath69 . [ fig : asf6:rt : x - band : rot],width=321 ] and g - shift of measured at @xmath25 ghz in the @xmath67 and @xmath70 planes . @xmath68 is the angle between the _ c _ axis and the external magnetic field @xmath69 [ fig : br : rt : x - band : rot],width=321 ] of measured at @xmath25 ghz in the principle directions . lines are results of linear fits . note that the discontinuity of the data at @xmath18 k is likely to be associated with the orthorhombic to monoclinic phase transition@xcite , and that the topology of the co structure in the low temperature monoclinic phase is not known yet . [ fig : asf6:lw : x - band : tempdep],width=321 ] we first discuss the low frequency limit at 300 k. the room temperature anisotropy of the line width , @xmath8 and the shift relative to the free electron g - value , @xmath35 , measured at @xmath25 ghz are shown in fig . [ fig : asf6:rt : x - band : rot ] for the asf@xmath19 compound and in fig . [ fig : br : rt : x - band : rot ] for the br compound . since the frequency dependence is small , the data at 9.4 ghz correspond to the @xmath71 limit . the simplest explanation is that most of the angular dependence of the g - factor reflects the g - factor anisotropy of the cation molecules . the angular dependence of @xmath35 is the same in the two systems , supporting that this is a property related to the cations . for a magnetic field in the _ a _ direction , spins are on equivalent sites with the same g - factor . here a `` site '' consists of a pair of neighboring high and low charge molecules along _ a_. since the high charge molecule carries most of the spin , the g - factor of the compound along _ a _ is close to that of a singly charged cation with field perpendicular to the molecular plane . in the _ b _ and _ c _ directions there are two sites with respect to the magnetic field . however , by symmetry , the g - factors of the two sites are the same for these particular directions . the g - factors are different in general directions , but the two esr lines of the isolated chains are `` motionally '' narrowed by the perpendicular exchange interaction between chains . the angular dependent parts of @xmath8 and @xmath35 are proportional within experimental accuracy . it would be tempting to assign the anisotropy of the line width to an incomplete motional narrowing of the inhomogeneity of the g - factors . however , as discussed below , the small frequency dependence of the line widths contradicts such an assignment , and we conclude that most of the zero frequency line width is from anisotropic exchange and dipolar coupling . the high temperature limit for the narrowing , @xmath72 , is of the order of @xmath73 and the measured line width is @xmath74 t . consequently @xmath75 t is sufficient to narrow lines . now we turn to the temperature dependence in the low frequency limit . the line width in the high - temperature orthorhombic phase is anisotropic and changes little with temperature for all frequencies and crystal orientations . @xmath8 has a small linearly temperature dependent contribution . on fig . [ fig : asf6:lw : x - band : tempdep ] we plot the temperature dependence of the line width of measured at 9.4 ghz frequency . the rate of increase of @xmath8 with temperature is @xmath76 in the three crystallographic orientations respectively . the low temperature linear limit does not apply since the exchange integral @xmath3 k determined from the bonner - fisher fits of the susceptibility ( fig . [ fig : asf6:br : susc ] ) is not much larger than the temperature . the experimental data fit with @xmath77 exchange anisotropy for the @xmath78 salt with the assumption that all the temperature dependence of the line width is described by eq . ( [ eq : anis ] ) . [ cols="^,^,^,^,^,^ " , ] k. measurements in the _ a _ , _ b _ and _ c _ orientations are plotted by solid circle , square and triangle respectively . lines show parabolic frequency dependence . a ) frequency dependent part of the line width . b ) frequency dependent part of the g - factor @xmath79 . [ fig : br : freq : temp : lw],width=321 ] we now concentrate on the frequency dependence of @xmath8 and the g - factor . fig . [ fig : br : freq : temp : lw ] shows the @xmath80 k data in . the solid lines in fig . [ fig : br : freq : temp : lw ] are fits to eq . [ eq1 ] and [ eq2 ] . clearly , both the g - value and @xmath8 have a quadratic frequency dependence in all three crystallographic orientations . the frequency and temperature dependence is similar in the asf@xmath0 compound ( not shown ) . the coefficients measured at @xmath80 , @xmath81 and 200 k are tabulated in table [ table:1 ] . the following arguments suggest that in these compounds the unusual quadratic frequency dependence arises from a staggered dm field . another mechanism could be an incomplete motional narrowing of a g - factor inhomogeneity@xcite . as discussed above , the g factor of sites is uniform within the chains and alternates only from chain to chain . the quadratic dependence of the line width is largest for magnetic fields in the _ a _ direction where the g - factor is obviously homogeneous . in fact , the g factors of all sites are the same for magnetic fields in all three principle directions . thus only in general directions can the g - factor anizotropy contribute to the frequency dependence of the line width . on the other hand , the dm interaction between neighboring sites results in an alternating effective field along _ a _ direction chains . the neighboring sites are not related by inversion symmetry and thus the dm interaction is non - zero . the dm interaction is also non zero between sites of neighboring chains in the _ ( a , b ) _ plane . although the quadratic frequency dependence of the line width is well explained by the dm mechanism , the temperature dependence of the quadratic term does not follow eq . ( [ eq : dmanish ] ) . the same applies to the quadratic field dependence of the g - factor which does not strengthen at lower temperatures as predicted by eq . ( [ eq : dmanisg ] ) . it is quite possible that the low temperature limit is no more valid above 200 k and a better approximation is given by the high temperature limit . an order of magnitude estimate of the magnitude of the dm vector , @xmath82 from the quadratic dependence of the line width , @xmath83 is @xmath84 t. the magnitude of @xmath85 is not unrealistic , it is an order of magnitude less than in another organic weak ferromagnet , @xcite , where the interaction is between singly charged pairs of et molecules . we presented an experimental study to test theoretical predictions for the temperature and magnetic field dependence of the esr spectrum of a quantum @xmath86 antiferromagnetic chain with small anisotropic and alternating interactions . we find that above 190 k the quasi one dimensional compounds have a @xmath2 heisenberg antiferromagnetic chain - like magnetic susceptibility with @xmath3 k and @xmath4 k coupling constants for x = asf@xmath0 and br respectively . from a weak temperature dependence of @xmath8 in the @xmath87 temperature range we estimate an exchange anisotropy @xmath88 in the asf@xmath89 compound . we argue that the unusual quadratic magnetic field dependence of the line width and g - factors arise from a dzyaloshinskii - moriya anisotropic exchange interaction . the dm interaction , observed here as a small anomaly in the esr parameters , has important implications for the low temperature antiferromagnetic phase : namely the material is likely to be a weak ferromagnet . earlier magnetization measurements@xcite report an increase of magnetization below @xmath90 k. this small ferromagnetic contribution was attributed to a small impurity phase@xcite , but it could also be from a weak ferromagnetism . we are grateful to richard gal for fruitful discussions and for his support in the experimental work at epfl . this work was supported by the swiss national science foundation and its nccr manep . work at budapest was supported by the hungarian state grants otka pf63954 , k68807 . n.b . acknowledges the support from the prospective research program no . pbelp2 - 125427 of the swiss nsf . work at angers was supported by the french national research agency interdisciplinary anr project @xmath91-filled 2009 - 2011 ( anr-08-blan-0140 - 01 ) and the cnrs .	we present a multi - frequency electron spin resonance ( esr ) study in the range of 4 ghz to 420 ghz of the quasi - one - dimensional , non - dimerized , quarter - filled mott insulators , ( x = asf@xmath0 , br ) . in the high temperature orthorhombic phase above @xmath1 k , the magnitude and the temperature dependence of the high temperature spin susceptibility are described by a @xmath2 heisenberg antiferromagnetic chain with @xmath3 k and @xmath4 k coupling constants for x = asf@xmath0 and br respectively . we estimate from the temperature dependence of the line width @xmath5 an exchange anisotropy , @xmath6 of @xmath7 . the frequency dependence of @xmath8 and the g - shift have an unusual quadratic dependence in all crystallographic orientations that we attribute to an antisymmetric exchange ( dzyaloshinskii moriya ) interaction .
the existence of a threshold for bound - state formation in the two - body problem is one of the basic differences between quantum and classical mechanics @xcite . it is also well known that the existence of this finite threshold strongly depends on the spatial dimension and on the presence ( and quantum statistics ) of other particles @xcite . even more striking consequences arise however when we consider the three - body problem . in this case , according to quantum mechanics , three - body bound states are favored in suitable parameter regimes with respect to two - body bound states . this interplay between two - body and three - body physics leads to the so - called efimov effect @xcite . although its theoretical prediction by v. efimov dates back to 1970 , the incredible experimental advances in ultracold gases , which allow to observe and study this physical scenario in dilute atomic clouds @xcite , have triggered in recent years a strong revival of interest @xcite . + + + . different bands are plotted with different color ( the trion band is green , the off - site trion band is blue , the dimer band is red and the free particle band is purple ) . for weak interactions bands are not separated and therefore colors are only formal . [ spectrum ] in this paper we address the problem of three - body bound state formation in a lattice for three fermions with different values of an internal degree of freedom . our calculations show that three fermions can form different bound states ( see fig . [ sketch ] ) . this is directly relevant for ultracold atomic gases , since the problem under investigation corresponds to loading three fermions of the same atomic species ( e.g. @xmath4 [ ] or @xmath5 [ ] ) in different hyperfine states into an optical lattice . similar systems have already been studied in current experiments without superimposing an optical lattice , though the experimental observation is hampered by three - body losses @xcite . remarkably , by loading the gas into an optical lattice the effect of the losses could be suppressed , as a large rate of on - site three - body loss can prevent coherent tunneling processes from populating any site with three particles @xcite . for three - body loss , this mechanism would provide an effective three - body hard - core constraint @xcite , which suppresses the actual loss events . the analogous problem in the absence of a lattice has been investigated in ref . [ ] , and more recently in ref . [ ] by using the functional renormalization group , where the authors found a very interesting evolution of the system properties for increasing attraction . indeed , in the three - dimensional case , three - body bound states appear for smaller values of the attraction than two - body bound states . therefore there is a range of parameters where the ground state is a three - body bound state and no two - body bound states are allowed . in the strong - coupling regime instead a two - body bound state turns out to be the stable solution . this peculiar phenomenon , which is consistent with the efimov picture , is clearly mirrored in the phase diagram of the corresponding many - body problem @xcite , though the presence of other fermions triggers pair formation and superfluidity also at weak - coupling . the presence of a lattice is known to induce important differences in the many - body phase diagram , especially at strong - coupling and in presence of three - body losses . indeed , although the overall picture of the competition between superfluid and trimer formation ( on - site trionic phase ) is similar to the continuum case , the latter is stable in the lattice at strong - coupling @xcite , unless the three - body loss rate @xmath6 is very large @xcite . in the continuum instead the trionic phase is bound to disappear for large values of the attraction @xcite . + + in order to better understand this point and the differences to the continuum case , in this paper we take a step back and focus on the three - body version of the problem investigated in ref . specifically , we consider the case where all the interactions between the species are symmetric in a one - dimensional chain ( @xmath7 ) and a two - dimensional square lattice ( @xmath8 ) . we address this problem using a numerical approach , combining exact - diagonalization ( ed ) with the lanczos algorithm @xcite , which allows us to obtain information about the low - energy sector of the spectrum for much larger system sizes than the full ed approach . moreover we also study the effect of including an onsite three - body hard - core constraint , which mimics the effect of a very large three - body loss rate @xmath6 in presence of a lattice , as explained in ref . [ ] . the paper is organized as follows . in the next section , we introduce the model and give details about the methods . then , in section [ sec:1d ] , we discuss the results for a one - dimensional chain , first considering the case without losses , then introducing the three - body constraint induced by losses . in section [ sec:2d ] we consider the same problem in the two - dimensional case , in order to understand what is the role of dimensionality in the formation of three - body bound states and in the competition between two- and three - body bound states . conclusions are drawn in section [ sec : conclusion ] . the system under investigation is described by the following single - band hubbard hamiltonian : @xmath9 where @xmath10 denotes a summation over neighboring lattice sites , @xmath11 is the fermionic creation operator with hyperfine spin @xmath12 at site @xmath13 , while @xmath14 is the corresponding number operator . @xmath15 is the hopping amplitude between nearest neighbor sites and assumed as energy unit , while @xmath16 is the on - site attractive interaction . in the context of ultracold gases this hamiltonian can be realized by loading three fermions in different hyperfine states into an optical lattice . under suitable conditions @xcite , the effect of higher lattice bands can be neglected . the case in which the interactions are symmetric corresponds to equal scattering lengths between different hyperfine states . in practice this can be realized by different hyperfine states of @xmath4 in the presence of a strong magnetic field , or by alkaline - earth atoms such as @xmath5 . the hamiltonian is then ( approximately ) invariant under @xmath1 rotations in the space of the hyperfine states . the three - body interaction term with @xmath17 ( hard - core constraint ) is introduced to take the effects of three - body losses in the strong loss regime into account according to refs . on the other hand @xmath18 corresponds to the case when three - body losses are negligible . a finite value of @xmath19 however , is not equivalent to the regime of moderate loss , where actual losses can not be avoided , and the effective hamiltonian description above fails . therefore we do not consider this situation here . by using an exact diagonalization approach we were able to calculate the full spectrum of the problem for system with up to @xmath20 lattice sites for the linear chain in @xmath7 and @xmath21 lattice sites for a two - dimensional square lattice . since , however , the size of the system addressed is relatively small ( especially for the @xmath8 case where the finite - size effects are expected to be more severe ) , we also used the lanczos algorithm@xcite with full reorthogonalization to calculate low - lying eigenvalues and eigenvectors for system sizes up to @xmath22 lattice sites . we also carefully studied the effect of closed and periodic boundary conditions . all the results shown in this paper are obtained with periodic boundary conditions , which preserve the translational invariance of the lattice and minimize finite - size effects . in this section we present numerical results for a one - dimensional chain . for the @xmath23-th eigenvalue of the system @xmath24 , the corresponding eigenstate can be expressed as @xmath25 provides us with the probability of having the particles with the hyperfine - state @xmath12 at the positions @xmath26 in the @xmath23-th energy level . since the system is translationally invariant , we can consider the following 3-particle probability distribution : @xmath27 which is only a function of the relative positions of the fermions with the hyperfine - state @xmath28 and @xmath29 with respect to the fermion with hyperfine - state @xmath30 . we can anticipate that our results will prove that only the long - distance behavior of the many - body wave - function ( and therefore of @xmath31 ) provides clear - cut signatures of bound - states formation . this is a crucial difference between the results presented in the next subsection and ref . [ ] , where only the energy spectrum and onsite observables were investigated , and also the main reason we arrive at somewhat different conclusions . we start by considering the case where three - body losses do not play any role , i.e. we take @xmath18 in eq . ( [ hamiltonian ] ) . first we study the evolution of the ground state properties of the system as a function of the interaction . as shown in fig . [ gs_wave_function ] , for all values of the interaction the ground - state probability distribution @xmath32 has a pronounced maximum at @xmath33 and a fast decay with the distance away from this point . we also found that , already for a relatively small system size , the tails of the probability distribution at large distances are essentially unaffected by the size of the lattice . their fast decay is consistent with the formation of three - body bound states ( trions ) for arbitrarily small interactions . therefore , similar to the situation for for two - body bound states in continuum space @xcite , there is no threshold for three - body bound states formation in @xmath7 . it is easy to realize that at weak - coupling their wavefunction has a finite spread in space over many lattice sites and therefore trions can be seen as local objects only for large values of the interaction . as a function of the eigenstate label @xmath23 for increasing energy and different values of the interaction : @xmath34 blue dashed line with squares , @xmath35 red solid line with circles . the data shown correspond to a linear chain of @xmath36,scaledwidth=45.0% ] plotted as a function of the interaction for different values of the system size : @xmath37 ( blue solid line with squares , red dashed line with circles and green dashed - dotted line with triangles respectively ) . the inset shows the same plot in log - scale ( y - axis ) , in order to underline the relevance of finite - size effects at weak - coupling.,scaledwidth=45.0% ] + + ) as a function of @xmath38 for @xmath39 . different bands are plotted with different color ( the off - site trion band is blue , the dimer band is red and the free particle band is purple ) . for weak interactions bands are not separated and therefore colors are only formal.,scaledwidth=45.0% ] the eigenvalue spectrum shown in fig . [ 1d_ns19_spectrum ] allows to distinguish three main structures which are clearly separated from each other only for strong interactions , while they overlap at weak - coupling . we found that the lowest band separates from the second band for @xmath40 in agreement with ref . moreover a further study of the low - energy spectrum using the lanczos method ( see fig . [ 1d_ns51_spectrum ] ) shows that this threshold converges very quickly as a function of the system size . we then characterize the eigenstates through a detailed study of @xmath41 . as one would expect , the analysis reveals that the states of the lowest band ( see fig . [ on - site - trions ] ) are trionic in nature and @xmath41 is maximal in the center . moreover , since trions are already essentially local objects for @xmath42 , this provides a simple explanation for the observed system size - independence of @xmath43 . more surprising is the existence of a fine structure inside the second band , which is not visible on the scale of the plot in fig . [ 1d_ns19_spectrum ] . indeed the second band is further split into three sub - bands which are separated by extremely small gaps at strong - coupling of the order of @xmath44 . despite being so close in energy these sub - bands accommodate states which are very different in nature . the existence of a fine structure in the spectrum and the asymptotic size of the gap are in agreement with the results found at strong - coupling for the analogous problem with three indistinguishable bosons@xcite , which also provides an insightful picture of the strong - coupling regime and further confirms our results . + + the central sub - band is composed of proper dimer states where the three - body bound states are dissociated . this is clearly evident in figs . [ dimer1 ] and [ dimer2 ] , where @xmath41 corresponding to different states in the dimer sub - band are shown . in these cases @xmath45 does not decay anymore at large distances and is maximal along the lines @xmath46 , @xmath47 or @xmath48 , being very small in the origin @xmath33 . these eigenstates correspond to configurations where a two - body bound state coexists with an unbound particle . the states below and above the dimer sub - band show instead a completely different behavior , since their probability distribution has a fast decay at large distances , as shown in figs . [ off - site - trions1 ] and [ off - site - trions2 ] . this indicates that they are excited three - body bound states . this indicates that they are excited three - body bound state , i.e. the third particle still stays bound to a dimer , although their energies are very close to dimer states . the probability distribution at @xmath33 is however substantially suppressed as in the dimer states , while is maximal when a pair sits in one of the nearest - neighbors sites of the third particle . we therefore refer to these states as off - site trions , in order to distinguish them from the states in the lowest band . for distinction we will refer from now on to the latter as on - site trions . our interpretation is that in the case of off - site trions the increase of local quantum fluctuations with increasing energy is strong enough to reduce the probability of finding three fermions on the same site @xmath49 basically to zero even at strong - coupling , while the three - body bound state character , which is related to the large - distance behavior of @xmath41 , is not changed . indeed the probability distribution of dimer states and off - site trions are pretty similar at short distances , and this probably also explains why they are so close in energy . we were not able to find simple arguments to explain the existence of off - site trions and also the size of the gaps between the off - site trionic and dimer sub - bands , which are for example absent in @xmath3 , as shown in sec . [ sec:2d ] . finally none of the states of the highest band in fig . [ 1d_ns19_spectrum ] are bound , and the probability distribution @xmath41 is essentially flat and small over the whole system , as evident in fig . [ particle - band ] . we refer to this structure as the unbound particles band . it is thus interesting to observe that having a maximum in the probability of finding three - particles in the same lattice site is not essential to three - body bound state formation , which is instead unambiguously related to the long distance behavior of the probability distribution . in order to better understand this point we further studied @xmath49 , i.e. the probability to find all three particles at the same lattice site . this parameter was used in ref . in order to identify trions and hence called _ trionic weight _ by the authors . from the considerations above it is however clear that the existence of off - site trions and the non - local character of the trionic wave - function at weak - coupling makes this interpretation inconclusive . indeed , the very irregular behavior of @xmath49 shown in fig . [ trionic_weight ] in the weak - coupling regime is simply due to the fact that the onsite trionic band overlaps with the higher bands . therefore some of the states shown in fig . [ trionic_weight ] are onsite trions , while others are still trions but off - site in character or just dimers . for small @xmath50 these states are very close in energy and only the detailed study of the eigenstates that we performed can distinguish between them . even though the states in fig . [ trionic_weight ] with a large @xmath49 are indeed on - site trions , a decrease in @xmath49 does not always correspond to an unbound state because of the presence of off - site trions . moreover , in the weak - coupling regime there are dimer or even specific unbound particle states which have a probability for triple occupation even larger than for three - body bound states . + + our point made above is dramatically evident at strong coupling , where the onsite trionic band is clearly separated in energy from the higher bands . indeed , while @xmath49 is close to one for all the states in the onsite trionic band , indicating that onsite trions are essentially local objects for this value of the coupling , the same parameter drops to a very small value for off - site trions , which are still three - body bound states . within the onsite trionic band , we also observe as in ref . [ ] that excited trions have a slightly higher value of @xmath49 , compared to the ground state . this is probably related to the higher potential energy contribution in the excited states and to the different relevance of quantum fluctuations . in conclusions local observables like @xmath49 are unable to distinguish between an off - site trionic and dimer states in all interaction regimes although their wave - functions are completely different , as shown above . therefore @xmath49 should in general not be taken as an indicator of three - body bound state formation . as shown in fig . [ gs_trionic_weight ] , we found the parameter @xmath51 in the ground state to increase smoothly with @xmath52 independently of the system size in contrast to ref . [ ] , where a step - like behavior was observed . in the strong - coupling limit we expect that trions behave as composite particles and it makes sense to investigate their dispersion relation . the results shown in fig . [ dispersion ] prove that trionic states are distributed in their band as free - particles in the tight - binding approximation . for onsite trions the effective chemical potential @xmath53 and hopping @xmath54 agree very well with perturbative estimates given in ref . [ ] . for off - site trions instead @xmath55 of the three particle system as a function of @xmath56 for different values of lattice size @xmath57 . the inset shows the ratio @xmath58 , where @xmath59 is linear size of the system.,scaledwidth=45.0% ] + as already mentioned in the introduction , realizations of three - component fermi system in ultracold atomic gases are plagued by three - body losses , which strongly limit the lifetime for observing the interesting phenomena under discussion here . however , as shown in refs . [ ] in the presence of a strong loss rate @xmath61 in a lattice , actual losses are suppressed and the system can be described by the hamiltonian in eq . [ hamiltonian ] provided one includes an effective three - body constraint ( @xmath17 ) . therefore , in order to understand the effect of a strong three - body loss rate on three - body bound state formation , we include the additional three - body term in the hamiltonian and set @xmath62 . as proven in ref . [ ] , this setup is indeed essentially indistinguishable from the real constraint . the energy spectrum of the system in the presence of the three - body constraint is shown in fig . [ spectrum_1d_constraint ] . due to the constraint , triple occupancies are forbidden and therefore onsite trions are suppressed ( i.e. their energy is lifted by an amount proportional to @xmath19 ) . one could then wonder if an onsite three - body constraint is sufficient to rule out the possibility of three - body bound states in the presence of an attractive two - body potential which is also onsite . the previous considerations on the existence of off - site trions clearly indicate that the probability of having triply occupied sites is not necessarily correlated with the formation of three - body bound states . our results fully confirm this expectation since in the constrained case we found the ground state of the system to be always a three - body bound state but off - site in nature . as before the off - site trionic states are separated by a small energy gap from the dimer states . since both off - site trionic and dimer states have similar values of @xmath49 already in the unconstrained case , the energetic hierarchy between them is only weakly affected by the constraint , despite of the suppression of triple occupancies . the existence of off - site trions and hence of an off - site trionic phase has been also observed in ref . [ ] , where the many - body version of the problem investigated here was addressed in one spatial dimension . therefore , at least in 1d , three - body bound state formation is a robust phenomenon which takes place for arbitrarily small values of the interaction and is robust against three - body losses . moreover , the presence of off - site trions show that quantum fluctuations can strongly suppress triple occupancies without necessarily inducing the breaking of trions in favor of dimers . in order to understand how much these conclusions are related to the peculiarity of the one - dimensional case , in the next section we will address the same problem in the context of a two - dimensional square lattice . as in the one - dimensional case , we first investigate the case where no losses are present and @xmath18 . the complete energy spectrum of the system on a square lattice with @xmath63 lattice site as a function of @xmath50 is shown in fig . [ 2d_ns4_spectrum ] . despite the same overall structure as in the one - dimensional case emerges from the picture , finite - size effects are clearly not - negligible . the fast growth of the hilbert space with the number @xmath64 of lattice sites strongly constrains the linear size of the lattice which we can handle within full ed and even within the lanczos algorithm we were only able to address up to @xmath65 sites ( see fig . [ 2d_bigger_spectrum ] ) . since finite - size effects are expected to scale with the dimension @xmath66 as a function of @xmath67 , it is clear that increasing dimensionality is extremely unfavorable for our numerical approach and a careful extrapolation to the thermodynamic limit is required for all observables . as in 1d the lowest band separates from the higher bands only beyond a critical coupling . the threshold value is roughly twice larger than the value in 1d ( @xmath68 for @xmath69 ) , suggesting that they could converge to the same value when rescaled with the bandwidth @xmath70 . the prefactor is however slightly bigger ( @xmath71 ) in the data obtained within the lanczos approach on a lattice with @xmath72 shown in fig . [ 2d_bigger_spectrum ] , suggesting that there could still be sizable finite - size effects affecting the threshold value . it is worth pointing out that the lowest band now follows rather closely the higher bands in a finite range of couplings , as evident in fig . [ 2d_bigger_spectrum ] . since this could suggest the existence of a finite threshold for three - body bound state formation which is absent in 1d , we also studied in detail the eigenvectors and the probability distribution @xmath73 in the two - dimensional case . by using again the translational invariance , we obtain the following reduced probability distribution : @xmath74 since , however , @xmath75 is still a function of four variables , we further reduce them by averaging the probability distribution @xmath76 over all sites at the same manhattan distance@xcite and define @xmath77 in the expression above the summation runs over @xmath78 and @xmath79 is the number of the lattice sites at a manhattan distance @xmath80 from the third - species fermion . while this allows us to produce meaningful three - dimensional plots also in 2d , it clearly averages out a part of the information contained in @xmath75 . moreover , we introduce the following quantity @xmath81 which measures the average diameter of the three - particle cluster , as clarified below . we first consider the evolution of ground state properties as a function of the interaction . in the strong - coupling regime shown in fig . [ on - site - trions_2d_b ] , where the bands in the spectrum are well separated , the probability distribution has a clear maximum at @xmath82 and a very fast decay with distance . in this regime the average size of the cluster @xmath83 plotted in fig . [ average_size ] is extremely small and essentially independent of the system size . thus we can confidently argue that at strong - coupling the ground state is an on - site trion . due to the relevance of finite - size effects it is much harder to draw analogous conclusions in the weak - coupling regime , as shown in [ on - site - trions_2d_a ] . indeed in this case the probability distribution stays finite at the edges of the lattice even for @xmath72 , as it was the case for dimer states in 1d . on the other hand , since also three - body bound states are expected to have a rather large spread of the wavefunction in the weak - coupling regime , the finite value found at the border is equally likely to be a residual finite - size effect . ) as a function of @xmath38 for @xmath84 . different bands are plotted with different color ( the off - site trion band is blue , the dimer band is red and the free particle band is purple ) . for weak interactions bands are not separated and therefore colors are only formal.,scaledwidth=45.0% ] in order to further clarify this issue we investigated the ratio @xmath58 as a function of the interaction strength for different values of @xmath64 ( see inset in fig . [ average_size ] ) . one would expect indeed a different scaling with the system size for trions and dimer states . our data clearly show that at strong - coupling @xmath58 scales to zero as expected for three - body bound states . at weak - coupling instead @xmath58 seems at first - sight to converge to a finite value , with a steep crossover between the two regimes around @xmath85 , where we identify @xmath86 as the inflection point of @xmath87 where @xmath88 . this threshold value clearly decreases in modulus with increasing system size , but we are unable to determine whether @xmath86 converges to zero or a finite value in the thermodynamic limit , due to the finiteness of our system . it is worth mentioning that @xmath89 seems to be compatible with a logarithmic scaling with the system size , as evident from fitting our numerical results . while this makes virtually impossible to draw a clear - cut conclusion only based on our exact - diagonalization method , it has instead a relevance for real systems . indeed this would imply the existence in finite 2d systems , such like e.g. in ultracold gases experiments , of a finite threshold for the crossover to a three - body bound state for all the practical purposes . similar arguments have already been used in the context of ultracold gases@xcite to explain the existence of quasi - condensates in bosonic systems in 2d , in analogy with the existence of a finite magnetization at finite temperature in the finite two - dimensional xy model @xcite . the study of excited states resulted in an outcome very similar to the 1d case ( see fig . [ exited_2d ] ) . the main difference compared to the one - dimensional case is that the fine structure within the second band in the energy spectrum is now absent and off - site trionic and dimer states are mixed . this is probably due to the larger bandwidth in 2d than in the one - dimensional case , which causes the different sub - bands to overlap . the lowest band still accommodates only on - site trions ( see fig . [ exited_on - site - trions_2d ] ) and the lowest state in the second band is still an off - site trion ( see fig . [ off - site - trions_2d ] ) , at least in the strong - coupling regime where the nature of the three - body bound state is evident in our data . finally we set @xmath62 and study the effect of the loss - induced constraint on the two - dimensional system . the picture which emerges ( see fig . [ spectrum_2d_constraint ] ) is pretty similar to the one - dimensional case and the lowest - energy off - site trion is promoted to the role of ground - state in presence of constraint . due to the greater delocalization of the off - site trionic wavefunction , it is even harder to assess the formation of three - body bound states at weak coupling in this case . in this work we addressed the problem of three fermions in different internal states loaded in a lattice , using an exact diagonalization approach . we studied the attractive and full @xmath1-symmetric case , focusing on the competition between three - body and two - body bound state formation . our results show that in one dimension three - body bound states ( trions ) are formed for arbitrarily small attraction between the fermions and they remain the lowest energy states for every value of the interaction . we also observed the existence of excited three - body bound states which share the same short - distance structure and typical energy of two - body bound states ( dimers ) , while they still feature the characteristic long - distance decay in the wavefunction of three - body bound states . the probability of triple occupancy is strongly suppressed in these states , in contrast to the `` canonical '' trionic states . we therefore refer to them as off - site trions in order to distinguish them from the states in the lowest energy band ( on - site trions ) . these results prove unambiguously that the knowledge of the spectrum and onsite observables alone , like in ref . [ ] , do not provide a clear - cut distinction among two - body and three - body bound - states and can thus lead to misleading conclusions . in this respect only the detailed study of the long - distance behavior of the 3-body eigenstates is able to provide a proper characterization of the problem . the introduction of a three - body constraint mimicking three - body losses in the strong - loss regime as in ref . [ ] , inhibits on - site trions formation and promotes off - site trions to the role of lowest energy states . in two dimensions , our data suggest a logarithmic scaling with the system size of the threshold interaction for three - body bound states formation . the greater relevance of finite - size effects in the two - dimensional case did not allow us to assess if there is a finite threshold in the thermodynamic limit , although the scaling we found implies the existence of a crossover at a finite value of the interaction for all practical purposes in finite systems , like e.g. in ultracold gases experiments . this also suggests that a finite threshold is likely to exist in the three - dimensional case , which we can not address at present using the methods in this paper and which we plan to address in a future work by using monte carlo techniques . this work was supported by the deutsche forschungsgemeinschaft ( dfg ) via sonderforschungbereich sfb / tr 49 and forschergruppe for 801 . ap thanks c. verdozzi and the division of mathematical physics at lund university for their hospitality during the completion of this work and m. capone for the financial support by the european research council under fp7/erc starting independent research grant superbad " ( grant agreement n. 240524 ) . it thanks sfb 925 , project b5 for financial support .	we study the problem of three ultracold fermions in different hyperfine states loaded into a lattice with spatial dimension @xmath0 . we consider @xmath1-symmetric attractive interactions and also eventually include a three - body constraint , which mimics the effect of three - body losses in the strong - loss regime . we combine exact diagonalization with the lanczos algorithm , and evaluate both the eigenvalues and the eigenstates of the problem . in @xmath2 , we find that the ground state is always a three - body bound state ( trion ) for arbitrarily small interaction , while in @xmath3 , due to the stronger influence of finite - size effects , we are not able to provide conclusive evidence of the existence of a finite threshold for trion formation . our data are however compatible with a threshold value which vanishes logarithmically with the size of the system . moreover we are able to identify the presence of a fine structure inside the spectrum , which is associated with off - site trionic states . the characterization of these states shows that only the long - distance behavior of the eigenstate wavefunctions provides clear - cut signatures about the nature of bound states and that onsite observables are not enough to discriminate between them . the inclusion of a three - body constraint due to losses promotes these off - site trions to the role of lowest energy states , at least in the strong - coupling regime .
the process of parametric down conversion , in which pump photons with frequency @xmath0 are sent through a nonlinear crystal and converted into pairs of signal and idler photons with frequencies @xmath1 and @xmath2 , is a well - known and widely used phenomenon in quantum optics . since the conversion efficiency is rather small , in the optical parametric oscillator ( opo ) the process is enhanced by placing the crystal inside an optical cavity . due to its importance as a source of squeezed light when operated below threshold , the opo has been subject of much investigation , in particular for the case of a time independent continuous - wave pump field @xcite . the case of a periodically modulated pump field has been studied in @xcite . in the standard treatment of the opo , the calculations are performed in frequency space , the pump is taken as a time independent monochromatic beam , and it is often assumed that the cavity allows only one or two modes with well - defined frequencies , depending on whether degenerate or nondegenerate operation is considered . the field leaking out of the cavity is then determined from input - output formalism . an opo may also be pumped with a pulsed light field . this case is often discussed as if the opo generates a single mode pulse of squeezed light , but this is not quite the case . the output consists of a number of independent modes squeezed by different amounts as discussed for single pass down conversion in refs.@xcite . in the present paper we use a completely different approach to characterize the output from a pulsed opo . the treatment is a generalization of the discussion of the continuous - wave opo in ref . @xcite , and in contrast to refs . @xcite all calculations are performed in the time domain . the time domain treatment has several advantages . firstly , it is not necessary to assume the existence of a single mode or a couple of independent modes in the cavity since all effects of cavity resonances appear naturally in the analysis from the specification of the length of the cavity via the cavity round trip time . furthermore , since we do not assume orthogonal cavity modes , our treatment is valid for all values of the transmission of the opo output coupling mirror and not just for small transmissions . in particular , we may increase the mirror transmission to unity in our formulas and obtain the results for single pass down conversion . secondly , we do not assume any particular temporal shape of the pump field , and the analysis is thus suitable to investigate the transition between the few mode situation for short pulses and the highly multimode regime for continuous - wave fields . in the limit of a time independent pump field our results reduce to those of ref.@xcite , provided we invoke extra approximations which guarantee the existence of only a single cavity mode . thirdly , the time domain analysis is convenient if the light emitted from the opo is detected continuously in time and the back action of the measurements on the system has to be taken into account , see @xcite . although the opo output can be both squeezed and entangled , it belongs to the family of so - called gaussian states , and for a number of applications in quantum communication and computing , non - gaussian states are necessary . the opo output can be transformed into non - gaussian states by measurements as shown schematically in fig.[setup ] : a small fraction of the light is extracted , possibly frequency filtered , and detected by an avalanche photo diode ( apd ) . as , e.g. , a squeezed vacuum state is a superposition of even photon number states , a photo detection in the apd heralds the generation of a state with only odd photon number contributions such as single photon states and odd schrdinger kitten states in the remaining beam as verified experimentally in @xcite . @xcite provides a theoretical determination of the mode of the output state with the largest single photon fidelity for the case of a continuous - wave nondegenerate opo . we note that in the experimental verification by homodyne detection , one may use a constant amplitude local oscillator , measure the homodyne detector signal as a function of time , and then at a later stage multiply the recorded data with the mode function , selected according to the instant of the photo detection by the apd @xcite . for several applications of the heralded non - gaussian field states it is , however , necessary to mode match the state in real time . this is the case if one wants to interfere different pulses and if one wants to store the field state in an atomic medium and , e.g. , perform quantum gates by optical non - linearities in the medium @xcite . whether making use of electromagnetically induced transparency @xcite , a raman transfer @xcite , or faraday polarization rotation @xcite , storage of light pulses in atomic media involves a strong pump pulse that defines the temporal mode of the quantized field to be mapped onto the atomic collective degree of freedom @xcite . this implies that storage of a field state heralded by an apd detection event requires a storage control field , which has to be produced at the appropriate time . since the mode function occupied by the non - gaussian state may extend to times both before and after the trigger event @xcite , it may thus be necessary to switch on the control field _ before _ the trigger event or , more realistically , to delay the arrival of the signal to the storage medium . for demonstration experiments , it would be attractive to apply the storage protocol irrespective of the apd output and subsequently post select the instances where apd detection actually occurred in a suitable time window around the instant for which the stored mode is optimal . in this paper we do not model the storage process itself , but we identify the mode functions that maximize the non - gaussian features of the pulses to be stored , conditioned on realistic detection events . in particular , we wish to identify mode functions that are optimal independent of the precise detection time within a predefined time interval , which can not be chosen arbitrarily small if one wants a satisfactory apd detection probability . we shall for concreteness take the optimal mode to be the one with the most negative value of the wigner function at the origin . the paper is structured as follows . in sec . ii we develop a theoretical description of the opo and derive the two - time correlation functions of the opo output field . in sec . iii we show that a treatment similar to the one applied to the opo can be used to express the output field from the filter in fig . [ setup ] in terms of the input field . in sec . iv we determine the mode of the conditioned state with the most negative value of the wigner function at the origin , we compute the probability to obtain a trigger detection event within a specified time interval , and we compare the performance of continuous - wave and pulsed operation of the opo for preparation of pulses that can be stored with predefined control pulses in an atomic medium . section v concludes the paper . a theoretical model of the opo is based on the setup illustrated in fig . we have chosen a four - sided ring cavity for mathematical convenience , but this is not essential to the analysis . the field annihilation operators of the input and the output fields are denoted @xmath3 and @xmath4 , respectively , while @xmath5 , @xmath6 , represents the field at different positions inside the cavity as shown , and @xmath7 is the annihilation operator of a field in the vacuum state . the beam splitter @xmath8 couples the input and output fields to the intra cavity field , and the fictitious beam splitter @xmath9 models losses in the system . we model the entire loss as if it takes place between the crystal in the upper part of the figure and the output coupling mirror , which represents a worst case situation . the crystal is pumped by a classical pump field @xmath10 , which is assumed to pass unhindered through the cavity mirrors . the parametric process in a non - linear crystal is described in @xcite , and leads , in the time domain , to the following mapping of the field incident on the crystal to the field leaving the crystal @xmath11 in @xmath12 , which is taken to be real , is proportional to the second order susceptibility of the crystal and to the length of the crystal . we have assumed that the bandwidth of the down conversion is infinite and neglected depletion of the pump field and differences in the phase matching conditions for down conversion to different frequencies . this is likely to be a good approximation in the experiment described in ref.@xcite , since the crystal inside the opo is only few millimeters long , and thousands of cavity modes are populated . also , modes far from the center of the spectrum are irrelevant because they are filtered out by other components in experiments ( see the next section ) . . the beam splitter @xmath8 couples the intra cavity field to the input ( vacuum ) field @xmath3 and the output field @xmath4 , and the beam splitter @xmath9 is included to model intra cavity losses . ] in order to express @xmath4 in terms of @xmath3 and @xmath7 , we use the beam splitter relations @xmath13 and express @xmath14 in terms of @xmath15 @xmath16 @xmath17 ( @xmath18 ) is the field transmission ( reflection ) coefficient of @xmath19 , and @xmath20 is the round trip time in the cavity . in , we have assumed that each cavity mirror induces a phase shift of @xmath21 , and we have defined the following quantities @xmath22 where @xmath23 is the traveling time from @xmath8 to the crystal . furthermore , for convenience , we have redefined @xmath7 according to @xmath24 , where @xmath25 is the traveling time from @xmath9 to @xmath8 . isolating @xmath4 from eqs . , , and and assuming @xmath26 for all @xmath27 , we find @xmath28 the requirement @xmath26 , which is satisfied for @xmath29 , where @xmath30 is an integer and @xmath31 is a constant , means that the successive squeezing operations add up in phase , i.e. , it is the same quadrature that is squeezed at times @xmath27 , @xmath32 , @xmath33 , @xmath34 . for the special case of a time independent continuous - wave pump field @xmath35 , below threshold @xmath36 , a fourier transform of leads to the following expression for the output field in frequency domain : @xmath37 where @xmath38 is half the frequency of the pump field , @xmath39 and @xmath40 assuming that the input field @xmath3 is in the vacuum state , the frequency correlation function @xmath41 shows that resonances occur for @xmath42 , @xmath43 , as expected . in particular , the degenerate frequency @xmath44 is only a resonance frequency if @xmath30 is even . to compare eq . to the results given in ref . @xcite for a cavity with a single resonance frequency at @xmath44 , we assume @xmath30 to be even and @xmath45 in eq . . in order to have well separated cavity modes we must also assume @xmath46 and @xmath47 , but to maintain a finite width of the modes as in ref . @xcite we keep @xmath48 and @xmath49 fixed . since @xmath50 and @xmath51 in this limit , the mean number of round trips in the cavity is large , and @xmath52 must be correspondingly small , i.e. , @xmath52 must be proportional to @xmath20 . for @xmath53 we obtain eq . 46 of ref.@xcite . returning to the general case , we compute the two - time correlation functions of the output field for @xmath3 in the vacuum state from eq . : @xmath54 and @xmath55 these two expressions and @xmath56 are sufficient to characterize the output state completely , because it is gaussian . the correlations of an arbitrary single mode with mode function @xmath57 and annihilation operator @xmath58 are easily calculated from eqs . and . we shall only be concerned with the degenerate case below , and we thus choose @xmath59 with @xmath60 real . the constant phase factor is included to obtain a real value of @xmath61 , which corresponds to the case , where the axes of the squeezing ellipse in phase space lie along the quadrature axes . in the following we take @xmath57 to be real and omit the factor @xmath62 in eq . . the variances of @xmath63 and @xmath64 are @xmath65 and @xmath66 , respectively , and the most efficiently squeezed mode is thus obtained by choosing @xmath57 as the eigenfunction with the smallest eigenvalue @xmath67 of an integral equation with either the integration kernel @xmath68 or @xmath69 , and @xmath67 is the corresponding variance . since the output field at time @xmath27 is only correlated to the output field at times @xmath70 , @xmath43 , the mode function of the most efficiently squeezed mode is in general a spike function that is only nonzero at times @xmath70 , @xmath71 , for some @xmath27 , but if the variations of the pump field take place on a time scale @xmath20 or more slowly , displaced spike functions , and therefore also a smooth mode function with the same envelope , will be only slightly less squeezed . in the limit of a time independent pump field , eqs . and reduce to @xmath72 and @xmath73 until now we have described the fields in terms of time dependent heisenberg picture operators , but it is also useful to consider the opo model from a schrdinger picture point of view . in particular , the latter approach is suitable for a numerical treatment of the pulsed opo . to this end we divide the light beams into small segments of ( infinitesimal ) duration @xmath74 and treat each segment as a single mode , i.e. , we define @xmath75 and replace the continuous field annihilation operator @xmath76 ( @xmath77 , @xmath4 , @xmath78 , or @xmath14 ) by the discrete annihilation operators @xmath79 localized at time @xmath17 . in practice , we will have to deal with a finite @xmath74 , and neglecting the field variation within each @xmath74 interval is equivalent to a cutoff in frequency , justified by the experimentally relevant frequency regime . a thorough discussion of continuous and discrete operator descriptions of light beams in both time and frequency space may be found in ref . since the output field and the cavity field are gaussian for a vacuum input field , the state of all the small light beam segments is efficiently represented by a wigner function . in general , the wigner function of an @xmath80-mode gaussian state with zero mean values is on the form @xmath81 where @xmath82 is a column vector of quadrature variables , and @xmath83 is the covariance matrix of the @xmath80 modes . @xmath84 , where @xmath85 , @xmath86 , @xmath87 , and @xmath88 is the field annihilation operator of mode @xmath89 . before the pump pulse reaches the crystal , all modes are in the vacuum state , and @xmath83 is the identity matrix . we include all the @xmath90 cavity modes and a sufficiently large number of @xmath3 and @xmath7 modes in . as time passes by modes are squeezed , when they hit the crystal , input and cavity modes are transformed into cavity and output modes , when they hit @xmath8 , and vacuum and cavity modes are transformed into cavity and lost modes , when they hit @xmath9 . for each of these transformations @xmath91 is transformed according to @xmath92 , where @xmath93 is a matrix , which is easily determined from eqs . , , and . the corresponding transformation of @xmath83 is @xmath94 . at the end of the calculation all rows and columns of @xmath83 that represent lost modes are erased , and the result is a matrix , which contains the same information as eqs . and . we have used this method to compute the results for the pulsed opo in sec . the opo produces pairwise quantum correlated fields in a large number of cavity field modes , and in experiments it is necessary to apply a frequency filter before the trigger detector , see fig.[setup ] , to ensure that the photon is derived from a well defined sideband , and that the signal field predominantly occupies a single mode . we now turn to a description of such a filter modeled by a cavity with two beam splitters @xmath8 and @xmath9 and two perfectly reflecting mirrors as depicted in fig . [ filter ] . copying the notation from sec . ii , @xmath3 is the input field , @xmath4 is the output field , @xmath7 is a field in the vacuum state , @xmath5 is the field at different positions inside the cavity ( see the figure ) , @xmath17 ( @xmath18 ) is the field transmission ( reflection ) coefficient of @xmath19 , and @xmath95 is the round trip time in the filter cavity . the fields are related according to the equations @xmath96 from which we derive @xmath97 transforming the output field to frequency space @xmath98 we find @xmath99 and it is apparent that the frequencies transmitted most efficiently through the filter are those which satisfy the condition @xmath100 , @xmath43 . the free spectral range of the filter is thus @xmath101 , and for a given @xmath95 the bandwidth is determined by @xmath102 and @xmath103 . is partially transmitted into the cavity through the beam splitter @xmath8 . the reflected component of the input field is lost as is the part of the intra cavity field that leaves the cavity through @xmath8 . the part of the intra cavity field that leaves the cavity through @xmath9 contributes to the output field @xmath4 . @xmath7 is a field in the vacuum state . ] in order to select just one of the frequency modes emerging from the opo , the bandwidth of the filter should be much smaller than the free spectral range of the opo , but larger than the bandwidth of the opo . also , the free spectral range of the filter should be large compared to the bandwidth of the parametric down conversion . experimentally it can be difficult to build cavities that satisfy the last condition , and one may use instead a sequence of different filter cavities @xcite . since we assumed an infinite bandwidth of the down conversion in sec . ii , we must , for consistency , use an infinitely small filter cavity in the theoretical treatment . we thus assume @xmath104 , @xmath46 , and @xmath105 for fixed @xmath106 and @xmath107 , where @xmath108 , @xmath43 , is a resonance frequency of the filter cavity , and expand eq . to lowest order : @xmath109 in this limit the filter transmission function is a lorentzian @xmath110 we consider a single mode @xmath57 of the output field , and define @xmath111 and @xmath112 according to @xmath113 since the vacuum state does not contribute to normally ordered expectation values , we shall not need @xmath112 in the following , but from eq . @xmath114 i.e. , the action of the filter is effectively to transform the mode function @xmath57 into the ( not properly normalized ) mode function @xmath111 . note that if the time dependent part of the phase of @xmath57 is chosen as @xmath115 , the time dependent part of the phase of @xmath111 is also given as @xmath115 . when we choose the resonance frequency of the filter to equal half the carrier frequency of the opo pump beam , we may thus continue to use real mode functions and omit the factor @xmath116 in eq . . for a given bandwidth @xmath117 of the filter , @xmath118 is maximal for @xmath119 , and we thus assume @xmath120 in the rest of the paper . with the necessary tools at hand we can now proceed to an analysis of the preparation of non - gaussian light states that can be stored in atomic samples . gaussian states have wigner functions that are positive for all arguments , and as a measure of non - gaussian character we shall refer to negative values of the wigner function occurring at the origin of phase space , e.g. , for odd number states and odd schrdinger cat states . in ref . @xcite we derived an expression for the value of the wigner function at the origin of the state of an arbitrary real mode of the multimode state generated when conditioning on a photo detection event @xmath121 @xmath122 are the elements of the gaussian covariance matrix of the mode in which the apd detection takes place ( quadrature variables 1 and 2 ) and the chosen mode of the output state ( quadrature variables 3 and 4 ) before conditioning . @xmath83 is computed from the definition given just below eq . by use of the mode functions of the two modes , the transformation of the trigger mode function due to the filter , eq . ( with @xmath4 replaced by the relevant linear combination of @xmath4 and the annihilation operator of the vacuum field entering into the system at the beam splitter in fig . [ setup ] ) , and the two - time correlation functions given in eqs . and . we take the trigger mode function to be constant in a time interval of duration @xmath123 positioned at time @xmath17 and zero otherwise , where @xmath123 is much shorter than all other time scales in the system . in appendix a we discuss the applied detector model and the choice of trigger mode function in more detail . as explained in the introduction the mode that will be stored is determined by the shape and timing of the strong storage pulse , and it is advantageous if one can initiate the generation of the storage pulse before it is known whether a trigger detection event will actually take place at the right time relative to the pulse stored . the protocol is thus probabilistic . we consider an attempted storage as successful if a trigger detection event takes place within a predefined time interval of duration @xmath124 , whose position on the time axis is determined relative to the pump pulse , if the opo is pulsed , and relative to the storage pulse , if the pump field is time independent . in practice , @xmath124 is at least as large as the temporal resolution of the apd detection system , which is of order 1 ns @xcite . since the shape of the storage pulse is independent of the actual trigger detection time , the quantity we optimize is the mean value of the wigner function at the origin @xmath125 where the sum is over the @xmath126 trigger modes inside the acceptance interval @xmath124 , @xmath127 is the probability to obtain a detection event in trigger mode @xmath89 , which for infinitesimal @xmath128 is equal to the expectation value of the number of photons in the @xmath89th trigger mode , and @xmath129 is the value of the wigner function at the origin of the chosen mode of the generated state conditioned on a detection in trigger mode @xmath89 given in eq .. the total trigger probability @xmath130 is small , and hence we neglect the possibility to have two or more detection events within the time interval @xmath124 . .parameters used in the numerical minimization of @xmath131 . @xmath132 includes both propagation efficiency in the trigger channel and the efficiency of the apd , while @xmath133 is the propagation efficiency of the signal channel . the values are chosen in accordance with the numbers given in refs.@xcite [ cols="<,^,^",options="header " , ] under realistic experimental conditions there will be significant losses , and these are included in the calculations presented below . the chosen experimental parameters are given in table [ values ] , and for the case of a pulsed opo we assume that the pump field has a gaussian envelope and write @xmath134 as @xmath135 optimizing for different values of @xmath136 , @xmath137 , and @xmath124 ( or @xmath52 and @xmath124 for the case of a time independent pump field ) by optimizing the shape of the mode function of the mode that is to be stored , we obtain the values of @xmath131 shown in fig . [ tnp]a , and the corresponding success probabilities are given in fig.[tnp]b . examples of optimized mode functions are plotted in figs . [ mode ] and [ formel ] . we note that the absolute temporal position of the interval @xmath124 has been chosen to maximize the success probability . also , the values of @xmath52 for the opo driven with a time independent pump field have been chosen in order to obtain a total flux of photons in the degenerate mode of the output field from the opo , which is of order @xmath138 as in ref . this corresponds to a success probability of @xmath139 for @xmath140 . the values of @xmath137 have been adjusted to obtain comparable success probabilities in the calculations for pulsed pump fields . c + and b ) success probability as a function of @xmath124 . solid lines : time independent continuous - wave ( cw ) pump field with @xmath141 , @xmath142 , @xmath143 , and @xmath144 ; dashed lines : pulsed pump field with @xmath145 and @xmath146 , @xmath147 , @xmath148 , and @xmath149 ; dotted lines : pulsed pump field with @xmath150 and @xmath151 , @xmath148 , and @xmath149 ; dash - dotted line : pulsed pump field with @xmath152 and @xmath153 ; dash - dotted line labeled sp : single pass ( sp ) operation for a pulsed pump field with @xmath150 and @xmath154 . within each series a smaller @xmath137 or @xmath52 corresponds to a more negative value of @xmath131 and a smaller success probability.,title="fig : " ] + + and b ) success probability as a function of @xmath124 . solid lines : time independent continuous - wave ( cw ) pump field with @xmath141 , @xmath142 , @xmath143 , and @xmath144 ; dashed lines : pulsed pump field with @xmath145 and @xmath146 , @xmath147 , @xmath148 , and @xmath149 ; dotted lines : pulsed pump field with @xmath150 and @xmath151 , @xmath148 , and @xmath149 ; dash - dotted line : pulsed pump field with @xmath152 and @xmath153 ; dash - dotted line labeled sp : single pass ( sp ) operation for a pulsed pump field with @xmath150 and @xmath154 . within each series a smaller @xmath137 or @xmath52 corresponds to a more negative value of @xmath131 and a smaller success probability.,title="fig : " ] considering the trends in fig [ tnp]b , we find that the success probability increases when @xmath52 increases and when @xmath137 increases for fixed @xmath136 as expected . the success probability is also an increasing function of @xmath124 , because more terms are included in the sum in eq . , when @xmath124 increases . for a time independent pump field the increase is linear , because all the @xmath127 s are equal , while the success probability levels off to a constant value for a pulsed pump field when @xmath124 increases beyond the temporal width of the intensity distribution of the generated output field , which is roughly of order the width of the intensity distribution of the pump pulse plus the mean lifetime of a photon in the opo cavity ( @xmath155 ) plus the mean lifetime of a photon in the filter cavity ( @xmath156 ) . the constant value is thus reached faster for short pulses . it is apparent from the figure that a decrease in @xmath136 must be accompanied by an increase in @xmath137 in order to keep the success probability unchanged . both of these changes increase the required peak value of @xmath52 , i.e. , a larger field strength or a larger conversion efficiency in the crystal is needed . to explain the trends in fig . [ tnp]a we note that @xmath131 of a state with density matrix @xmath157 is @xmath158 , and thus an optimization of @xmath131 is equivalent to a maximization of the odd photon number components . an ideal single mode squeezed vacuum state is a superposition of even photon number states , and if a photon is annihilated from the state , these are converted into odd photon number states . the present experiment deviates from this ideal situation in two respects . firstly , the generated state is a multimode state , since all the generated photon pairs do not belong to the same mode . the result is that the overlap between the optimal mode and the modes of the generated photon pairs is only partial , and this introduces even photon number components into the state of the optimal mode . the mechanism is particularly severe if photon pairs are generated in the outer regions of the optimal mode such that the overlap between the optimal mode and the mode of the photon pair is significantly smaller than unity , but also significantly larger than zero . this suggests that a short pump pulse , where the down conversion process is only turned on in a short period , will lead to a more negative value of @xmath131 , and this is also what is observed in fig . the effect is particularly pronounced for very large @xmath124 . in this limit @xmath131 approaches a constant value for pulsed pump fields , because @xmath127 is small for trigger modes far from the center of the intensity distribution of the generated state and these modes do thus not contribute significantly to the sums in eqs . and . for a time independent pump field , on the other hand , @xmath131 approaches the value of the unconditional state , because the success probability approaches unity for very large @xmath124 , and all storage attempts are accepted as successful . the second deviation from ideal behavior is that losses degrade the odd photon number states into both odd and even photon number states , and since a photon number state with a large number of photons is more fragile than a single - photon state , it is to be expected that a large intensity of down converted photons will lead to a less negative value of @xmath131 . short pulses must be intense in order to keep the success probability close to @xmath159 , and this explains why we do not obtain more negative values of @xmath131 than those presented in fig . [ tnp]a if the pulse duration is decreased below @xmath160 but rather observe less negative values for very short pulses . in fig . [ tnp ] we have included an example of single pass down conversion for @xmath161 , i.e. , we have increased @xmath162 to unity without changing the rest of the parameters in table [ values ] . when the opo cavity is absent , the temporal width of the light pulse is only broadened in the filter cavity , and this broadening is small since the mean lifetime of a photon in this cavity is only @xmath163 . it is thus possible to determine from the precise apd photo detection time whether the photon pair that gives rise to the apd detection was generated in the beginning , in the middle , or in the end of the pulse . for small @xmath124 this leads to a negative value of @xmath131 because we know rather precisely where the second photon in the pair is , if it has not been lost . on the other hand , if we do not discriminate between early and late trigger detection events a single optimal compromise will only have a small overlap with the actual modes , which are well localized in time . the resulting rapid increase towards positive values in @xmath131 with increasing @xmath124 can be avoided by increasing the mean lifetime of a photon in the filter cavity , corresponding to a smaller frequency width of the filter , and thus the intensity of the light transmitted through the filter will decrease , which leads to an even smaller success probability than in fig . the success probability can be increased by increasing @xmath137 , but this will also result in a less negative value of @xmath131 , and from these arguments it is preferable to apply pulsed fields and the opo cavity in the setup to get the most efficient generation of non - gaussian states for storage . the values of @xmath131 presented in fig . [ tnp]a are significantly above the theoretical minimum of @xmath164 , and we note that the results are all obtained for physical parameters , which are already achieved in experiments . more negative values of @xmath131 can be reached if it is possible to reduce the losses . if , for instance , we increase the signal channel transmission from @xmath165 to unity , @xmath131 is decreased to @xmath166 for @xmath161 , @xmath154 , and large @xmath124 , while the success probability is unchanged , and if we further assume zero loss in the cavity and reduce @xmath167 to @xmath168 , we find @xmath169 . an increase in the trigger channel transmission or in the apd detector efficiency will increase the success probability , and this increase can be transformed into a more negative value of @xmath131 by decreasing @xmath52 , @xmath137 , @xmath124 , or @xmath167 . lr a ) & + + b ) & + for a pulsed pump field with @xmath170 and @xmath154 . @xmath171 is chosen to coincide with the time where @xmath52 is maximal . a ) @xmath172 ( the inset is an enlarged view ) , b ) @xmath173 , and c ) @xmath140.,title="fig : " ] & for a pulsed pump field with @xmath170 and @xmath154 . @xmath171 is chosen to coincide with the time where @xmath52 is maximal . a ) @xmath172 ( the inset is an enlarged view ) , b ) @xmath173 , and c ) @xmath140.,title="fig : " ] considering the optimized mode functions for a pulsed pump field in fig . [ mode ] , it is apparent that a fast variation in the mode function is present for small values of @xmath124 , but this variation is smoothed out for larger @xmath124 . the inset in fig . [ mode]a shows that the period of the fast variation is @xmath20 , i.e. , the round trip time in the opo cavity . a photon generated in one of the time localized modes in the opo cavity can only leave the cavity at times separated by an integer number of @xmath20 s , and thus , in the output field from the opo , the two photons in a pair are separated by an integer number of @xmath20 s . the fast variation in the optimal mode function can thus be explained from the fact that the mean lifetime of a photon in the filter cavity is smaller than @xmath20 , since this means that the observed photon in a pair is typically delayed less than @xmath20 in the filter cavity , and the detection time provides partial information on the time of generation of the photon pair . the information is erased as @xmath124 approaches @xmath20 , and this leads to the steep rise in @xmath131 towards less negative values observed for all curves to the very left in fig . [ tnp]a . if the duration of the pump pulse is short compared to @xmath20 , the optimal mode function differs qualitatively from those presented in fig . [ mode ] . in this case the optimal mode function is a series of spikes separated by @xmath20 , and the width of the spikes is determined by the width of the pump pulse provided the width of the pump pulse is also short compared to @xmath124 . in the opposite limit of a time independent pump field , the optimal mode function for short @xmath124 is qualitatively a function , which has a maximum close to the time of the trigger detection and decays exponentially when moving to the right or left from this maximum . superimposed on this is a fast variation with period @xmath20 . the mode function is not completely symmetric around the time of the trigger detection due to the filtering . for larger values of @xmath124 the fast variation disappears , and the maximum becomes more rounded . ( solid line ) for @xmath150 , @xmath154 , and @xmath174 . the dotted line is the function in eq . , while the approximation ( dashed line ) is almost indistinguishable from the optimum mode ( solid line ) in the figure . @xmath175 for the dotted line , @xmath176 for the dashed line , and @xmath177 for the solid line . ] in fig . [ formel ] we show the temporal shape of the pump field and the optimal mode function for @xmath174 in the same graph , and it is apparent that the generated output field is both broadened and delayed due to the nonzero lifetime of a photon in the opo cavity as expected . since @xmath131 and @xmath178 are both independent of @xmath124 for large @xmath124 , and since we have to choose the same mode function independent of the precise time for the apd detection , the optimal mode function is determined solely from the generated output field when @xmath124 is large , and it seems reasonable that the optimal mode function just follows the square root of the intensity distribution of the generated state . without applying the above theory , a crude guess for the optimal mode function is @xmath179 and this function , which is also shown in the figure , is actually quite close to optimal . with the above theory at hand , however , it is easy to calculate the actual intensity distribution from eq . , and this leads to the mode function @xmath180 which is even closer to the optimal mode function . due to its simplicity , eq . can be useful in a storage experiment , where @xmath124 must be sufficiently large to obtain a satisfactory success probability . we note , however , that neither eq . nor eq . is able to reproduce the optimal mode function in fig . [ mode]a for @xmath172 . in conclusion , we have presented a multimode description in time domain of the optical parametric oscillator , which is valid for both pulsed and continuous - wave pump fields , and we have used the theory to analyze non - gaussian states that can be stored in atomic samples . so far light storage has been demonstrated with gaussian states only , and the storage and retrieval of a non - gaussian state is a hall mark in the demonstration of atom - light quantum interfaces . our analysis suggests that , with similar success probabilities , pulsed pump fields lead to more negative values of the wigner function of the stored state than time independent continuous - wave pump fields . an essential ingredient in the experiment is the temporal spreading induced by the opo and filter cavities , because it separates the two photons in a generated photon pair temporally by an unknown amount and ensure that it is impossible to infer the precise position of the second photon in a pair from the time at which a detection took place in the apd . without this spreading we do not obtain large negative values of @xmath131 for large @xmath124 . after having thus characterized the system and identified the optimum strategy for given setup and pulse parameters , one may now move a step further and try to design filters and pump pulses in order to generate optimal mode functions which have high non - gaussian state content and which are particularly easy to handle in the storage part of an experiment . in the present treatment we have chosen to use the negative value attained by the wigner function as a measure of the non - classical character of the state generated , but we note that only minor changes are required in order to optimize other features as for example the single - photon or schrdinger kitten state fidelity . the authors acknowledge discussions with jonas s.neergaard-nielsen , anders s. srensen , and eugene s. polzik . various detector models are discussed in the literature . considering first detection in a single mode , a detector may for instance be a photon number resolving detector or an on / off detector . the photon number resolving detector registers the number of photons in the detected mode , and the detected mode is projected onto the detected photon number state , while the on / off detector only distinguishes between the outcomes vacuum and not vacuum , and , depending on the measurement result , the detected mode is projected onto one of these subspaces . if a continuous beam of light is observed , a third detector model is often assumed , where a detection event at time @xmath27 corresponds to an annihilation of a photon in a mode of infinitesimal temporal width positioned at time @xmath27 . is derived under the assumption that a detection event is equivalent to application of the annihilation operator of the observed mode to the state of the system , but since the flux of trigger photons is small in the experiment considered in sec . iv and the possibility to have two or more photons in a trigger mode can be neglected , all three detector models lead to practically identical results . a real detector consists of a medium , which absorbs photons , and typically several different microscopic transitions in the medium are allowed . the detector is then able to detect photons in more than a single mode . in an apd a photon absorption is accompanied by an excitation of an electron from a bound state to a free or a solid state conduction band state , the signal is amplified , and the resulting photo current is measured . the information concerning the precise microscopic transition is lost in the amplification process , and , when we condition on the macroscopic detection event , we must average over all modes that could have led to the observed detection ( see @xcite ) . since the filter in front of the detector selects a narrow band of frequencies , we assume that the detector is equally sensitive to all frequencies transmitted by the filter , i.e. , the detector bandwidth is assumed to be infinitely broad compared to the filter bandwidth , and all modes are observed by the detector . in this case we should sum over a complete set of modes on the interval @xmath124 in eq . . the obtained values of @xmath131 and @xmath178 are independent of the choice of basis which may be seen from the following argument . assume that the mode functions @xmath181 and the mode functions @xmath182 both constitute a complete basis on the time interval @xmath124 . @xmath181 and @xmath182 are related by a unitary transformation @xmath183 , which implies that the corresponding mode annihilation operators @xmath184 and @xmath185 are related by @xmath186 . if a photon is annihilated in the mode @xmath187 , the density matrix is transformed according to @xmath188 , where tr denotes the trace . to perform the averaging we multiply by the probability @xmath189 to annihilate a photon in the mode @xmath187 , and sum over all @xmath89 . after normalization this leads to @xmath190 since @xmath191 , the transformed density matrix is independent of the choice of basis , and the wigner function is then also independent of the choice of basis . it is thus not necessary to use the modes , in which the detections actually take place , and for mathematical convenience , we have chosen to sum over time localized modes in eq . . 99 m. j. collett and c. w. gardiner , phys . a * 30 * , 1386 ( 1984 ) . p. d. drummond and m. d. reid , phys . a * 41 * , 3930 ( 1990 ) . y. j. lu and z. y. ou , phys . a * 62 * , 033804 ( 2000 ) . h. h. adamyan and g. yu . kryuchkyan , phys . a * 74 * , 023810 ( 2006 ) . r. s. bennink and r. w. boyd , phys . a * 66 * , 053815 ( 2002 ) . w. wasilewski , a. i. lvovsky , k. banaszek , and c. radzewicz , phys . 063819 ( 2006 ) . v. petersen , l. b. madsen , and k. mlmer , phys . a * 72 * , 053812 ( 2005 ) . a. ourjoumtsev , r. tualle - brouri , j. laurat , and p. grangier , science * 312 * , 83 ( 2006 ) . j. s. neergaard - nielsen , b. m. nielsen , c. hettich , k. mlmer , and e. s. polzik , phys . * 97 * , 083604 ( 2006 ) . k. wakui , h. takahashi , a. furusawa , and m. sasaki , opt . express * 15 * , 3568 ( 2007 ) . a. e. b. nielsen and k. mlmer , phys a * 75 * , 023806 ( 2007 ) . a. andr , m. bajcsy , a. s. zibrov , and m. d. lukin , phys . lett . * 94 * , 063902 ( 2005 ) . s. e. harris , phys . today * 50 * , no . 7 , 36 ( 1997 ) . a. e. kozhekin , k. mlmer , and e. polzik , phys . a * 62 * , 033809 ( 2000 ) . b. julsgaard , j. sherson , j. i. cirac , j. fiurek , and e. s. polzik , nature * 432 * , 482 ( 2004 ) . j. sherson , a. s. srensen , j. fiurek , k. mlmer , and e. s. polzik , phys . a * 74 * , 011802(r ) ( 2006 ) . i. novikova , a. v. gorshkov , d. f. phillips , a. s. srensen , m. d. lukin , and r. l. walsworth , phys . lett . * 98 * , 243602 ( 2007 ) . k. j. blow , r. loudon , s. j. d. phoenix , and t. j. shepherd , phys . a * 42 * , 4102 ( 1990 ) . j. s. neergaard - nielsen , b. m. nielsen , h. takahashi , a. i. vistnes , and e. s. polzik , opt . express * 15 * , 7940 ( 2007 ) . p. p. rohde and t. c. ralph , j. mod . opt . * 53 * , 1589 ( 2006 ) .	we present a multimode treatment of the optical parametric oscillator , which is valid for both pulsed and continuous - wave pump fields . the two - time correlation functions of the output field are derived , and we apply the theory to analyze a scheme for heralded production of non - classical field states that may be subsequently stored in an atomic quantum memory .
dust emission is an excellent tracer of mass within star - forming regions . unfortunately , deriving the mass from measurements of dust continuum emission is not straightforward . in particular , the emissivities of the dust grains themselves are highly uncertain . generally , dust emissivity is assumed to have a power - law form at millimeter wavelengths of @xmath3 , and uncertainties in the emissivity spectral index , @xmath1 , and the relative dust opacity , @xmath4 , can result in derived masses that are uncertain by factors of a few ( e.g. , * ? ? ? the emissivity spectral index of dust is a difficult quantity to measure . for instance , the equation relating mass and the observed emission ( see equation [ fluxeq ] below ) relies on the dust being isothermal , which is not a valid assumption in any real star - forming region . furthermore , when measuring dust emission near the peak of the sed , variations in temperature along the line of sight and noise in the continuum maps can make it difficult to derive accurate values of @xmath1 ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? in particular , traditional @xmath5 minimization techniques are susceptible to the aforementioned problems , whereas bayesian techniques can recover a more accurate value of @xmath1 from emission maps near the peak of the dust sed @xcite . the emissivity spectral index has been constrained observationally in a few nearby molecular clouds and cores . on small ( @xmath60.1pc ) scales , @xmath1 has been measured in l1448 ( @xmath7 , 450 @xmath8 850 ; * ? ? ? * ) and tmc-1c ( @xmath9 , 160 @xmath10 2100 ; * ? ? ? recent measurements of the emissivity spectral index in small samples of class 0 protostars find @xmath11 ( 862 @xmath8 3.3 mm ; * ? ? ? * ) and @xmath12 ( 1.3 mm @xmath10 2.7 mm ; * ? ? ? the emissivity spectral index has also been measured towards a few higher mass and more evolved objects , such as hot molecular cores ( @xmath13 , 800 @xmath8 870 ; * ? ? ? * ) and ultracompact hii regions ( @xmath14 , 60 @xmath8 1300 ; * ? ? ? * ) . on larger scales , i.e. , those of filaments and molecular clouds , the emissivity spectral index has been measured towards the orion nebula cluster ( @xmath15 , 450 @xmath10 1100 ; * ? ? ? * ) and perseus b1 ( @xmath16 , 160 @xmath8 850 ; * ? ? ? * ) , both of which exhibit a flatter spectral index towards the densest regions in the clouds . on the scale of an entire galaxy , @xcite used measurements between 70 @xmath8 500 to find that the emissivity spectral index varies from @xmath17 in the center of m33 to @xmath18 towards the edge of the galaxy . the emissivity spectral index of dust measured in the laboratory setting has a similarly large range ( @xmath19 ) , with some materials showing a temperature - dependent value of @xmath1 ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? the emissivity spectral index is claimed to be anticorrelated with temperature in the orion , @xmath20 ophiuchus and taurus molecular clouds ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , though this may be a result of line - of - sight temperature variations and uncertainties in the observed fluxes rather than representing a true property of the dust grains @xcite . the omc-2/3 region , at a distance of 414pc @xcite , is the richest star - forming filament within 500pc . accordingly , it has been studied intensively with instruments like scuba @xcite , mambo @xcite , and spitzer @xcite . these datasets have mapped the extent of the dust emission in omc-2/3 , found dense cores within the filament , and determined which cores harbour protostars @xcite . in this paper , we present new 3 mm continuum observations of the orion molecular cloud ( omc ) 2/3 region . we combine our new mustang 3 mm observations with a previously published 1.2 mm map @xcite and a previously published gas temperature map @xcite to derive the emissivity spectral index of dust in the filaments and cores of the omc-2/3 region . the long wavelengths in this analysis make the derived spectral index much less dependent on the dust temperature than analyses of data at wavelengths around the peak of the dust sed , substantially reducing a major source of uncertainty in most other studies . here we describe our new 3.3 mm continuum map and the previously published 1.2 mm continuum and nh@xmath0 maps of omc-2/3 . the 3.3 mm continuum emission from omc-2/3 was mapped with the robert c. byrd green bank telescope ( gbt ) telescope in green bank , west virginia , using the mustang array @xcite . data were acquired in three observing sessions from november 2010 to february 2012 and were reduced and calibrated using procedures described as outlined in @xcite . data were visually inspected and @xmath21 were rejected , mostly due to poor weather . after flagging problematic data , there were approximately 14 hours of integration time spent on source . flux calibration was accomplished by periodically measuring the flux density of @xmath22 ori ( betelgeuse ) , which was determined with reference to uranus assuming @xmath23 @xcite . a key step in the data reduction is the so - called `` common mode '' subtraction , consisting of subtracting an average of all good detectors signals at a given point in time from each detector s signal . this is necessary to remove signals due to low - level fluctuations in atmospheric emission , but has the side effect of removing astronomical signals larger than approximately the mustang camera s instantaneous field of view . a key difference between the data reduction procedures of @xcite and those here is that the emission in omc-2/3 is considerably brighter than the galaxy cluster sunyaev - zeldovich effect being studied in the former work . it is therefore feasible to reduce the amount of flux lost in extended sources arising from common mode subtraction by iterating between the map and timestream model domains , using the map estimated at the previous step to correct the timestream ( common mode ) model , as described in @xcite . spatial scales up to 2.4 arcminutes are recovered by the reduction and calibration procedures , and the noise in the resultant map is @xmath60.36mjy / beam . the gbt 3 mm beam can be characterized by a gaussian main beam component ( full - width at half - maximum ( fwhm ) of 8.6 and an amplitude of 0.94 ) and a gaussian error beam component ( fwhm = 27.6 and amplitude of 0.06 ) . the relative contribution of the gbt error beam on the 3 mm map is roughly 20% , with a maximum possible of uncertainty of 30% in some locations . as a result , the 3 mm map overestimates the flux found in an assumed 8.6 beam by about 20% , and we correct for this effect by multiplying our 3 mm maps by a factor of 0.8 . the resulting map , smoothed to 10.8 , is shown in figure [ fluxfig ] . the 1.2 mm continuum emission from omc-2/3 was mapped with the iram 30 m telescope at pico veleta , spain , using the mambo array . data were acquired between 1999 and 2002 and were reduced using the mopsi package following standard reduction steps . see @xcite for a more detailed description of the observations and analysis . the spatial scales recovered in the mambo map are greater than those of the mustang map , a complexity that is discussed in section [ filtering ] . the noise in the mambo map is @xmath68mjybeam@xmath24 . the iram 30 m beam has been characterized at 1.3 mm by a gaussian main beam ( fwhm = 10.5 and amplitude 0.975 ) , a gaussian first error beam ( fwhm = 125 and amplitude 0.005 ) and gaussian second and third error beams with larger fwhm and lower amplitudes @xcite . because the iram 30 m error beams have such small amplitudes relative to the main beam , we do not modify the 1.2 mm intensities as was done for the 3 mm map . the observed 1.2 mm emission map is also shown in figure [ fluxfig ] . the 3 mm and 1 mm emission maps look qualitatively very similar , as expected if thermal dust emission on the rayleigh - jeans tail is the dominant emission mechanism . the nh@xmath0 ( 1,1 ) and ( 2,2 ) lines were observed towards omc-2/3 by the very large array ( vla ) and the green bank telescope ( gbt ) . details of the observations and calibration are given in @xcite . the combined vla and gbt map has a spatial resolution of @xmath65 and a spectral resolution of 0.6 . the kinetic temperature of the gas was derived in @xcite following the procedures outlined in @xcite and @xcite . in this paper , the temperature at each position is determined from a smoothed and regridded @xmath25 map from @xcite , putting it on the same @xmath611 resolution grid of the 1.2 mm and 3.3 mm continuum maps . any position included in the continuum maps but not included in the nh@xmath0 maps was assigned the median temperature value of the map ( 16.5k ) . the range of temperatures in this analysis ( @xmath26 ) is relatively narrow , so an assumed value of 16.5k will not be too far off the true line - of - sight average temperature for any region in omc-2/3 . roughly 4% of the pixels in our analysis were assigned the median temperature value of 16.5k . for our analysis , we assume that the gas and dust temperatues are coupled , a valid assumption for densities greater than 10@xmath27@xmath28 @xcite . the dust temperature map thus derived is shown in figure [ tempbetafig ] . ground - based bolometer maps all suffer from some degree of spatial filtering as part of the data reduction process , to remove significant atmospheric emission . the mustang 3 mm map recovers less extended emission than the mambo 1 mm map , with spatial scales greater than @xmath62.4 filtered out . we therefore pass the mambo data through the the mustang pipeline because the former have sensitivity to emission on larger scales than the latter . this operation ensures that the same spatial filtering are applied to both the 1 mm and the 3 mm observations . this technique is similar to that employed by @xcite to compare _ herschel _ and jcmt data . as a result of this filtering , the spectral indices determined in this paper are derived for the smaller - scale material in the omc-2/3 filament and cores and are not indicative of the more diffuse material in the orion molecular cloud . in the relatively dense material traced by the mustang observations , the gas and dust temperatures are expected to be well coupled ( see section [ ammonia ] ) . we smooth the mustang 3 mm and nh@xmath0 temperature observations to the mambo resolution , then we place all three maps on a common grid . we consider only those pixels with signal to noise greater than 10 in both the mambo map and the mustang map . we assume that the mm - wavelength emission observed along the line - of - sight as a function of frequency is given by a modified blackbody , i.e. @xmath29 where @xmath30 and @xmath31 in equation [ fluxeq ] , @xmath32 is the flux density per beam , @xmath33 is the solid angle of the beam , @xmath34 is the blackbody emission from the dust at temperature @xmath35 , @xmath36 is the mean molecular weight of interstellar material in a molecular cloud per hydrogen molecule , @xmath37 is the mass of the hydrogen atom , @xmath38 is the column density of hydrogen molecules , and a gas - to - dust ratio of 100 is assumed . in equation ( [ kappaeq ] ) , @xmath39 @xmath40 g@xmath24 is the emissivity at 230ghz of the dust grains at a gas density of 10@xmath41 @xmath28 covered by a thin ice mantle ( * ? ? ? * column 6 of table 1 ) and @xmath1 is the emissivity spectral index of the dust . equation [ fluxeq ] assumes that the dust emission is optically thin , which may be a concern towards the peak of the dust spectral energy distribution ( sed ) but is of no concern for observations on the rayleigh - jeans tail of the sed . for instance , assuming standard dust to gas ratios and dust properties @xcite , the 1.2 mm emission from a dense core with a @xmath42-band extinction of @xmath43 would have an optical depth of 0.004 , while one would expect an optical depth of 0.1 for 350 emission . the assumption of optically thin emission at 1 mm and 3 mm is justified for the observations presented here . the main assumptions required for equation [ fluxeq ] to reflect accurately the emission as a function of frequency are 1 ) a lack of free - free emission , 2 ) a lack of anomalous microwave emission , and 3 ) a weak temperature dependence . we discuss these different cases in section [ discussion ] . the emissivity spectral index can be determined from the ratio of the mambo and mustang emission maps . the spectral index at each position in the map is given by : @xmath44 with @xmath32 and @xmath34 as defined in section [ analysis ] . we assume that the dust temperature is equal to the gas temperature ( see section [ rj ] ; * ? ? ? the resultant emissivity spectral map is shown in figure [ tempbetafig ] . the omc-2/3 region has a complex background of filamentary structures that are not filtered - out in our data ( see figure [ fluxfig ] ) . as such , source extraction algorithms that use intensity maxima and minima to identify structures ( e.g. , clumpfind ; see * ? ? ? * ) will include emission from these filaments in their extractions . therefore , we identified sources using the _ getsources _ algorithm , which includes a prescription for filament extraction in addition to compact structures ( see * ? ? ? * for details ) . briefly , _ getsources _ uses multiple spatial decompositions to identify structure over many scales and characterize source masks , which are defined by two dimensional gaussians . information from the source masks ( and filament masks ) are combined over each wavelength to improve the final source footprint ( i.e. , in the case of blended sources at lower resolution ) and to extract source properties . for efficiency , we masked out the edges of the 1 mm and 3 mm maps to avoid noisy regions ( and false detections ) associated with the low - coverage areas . using _ getsources _ version 1.140127 , we identified initially 42 sources . this initial catalogue may include unreliable detections , e.g. , due to the filtering . therefore , we required all sources to be well detected in the single - scale decompositions ( i.e. , with significance @xmath45 1 ) . of the initial 42 sources , we rejected 12 for having either a low global significance or a low significances in either the 1 mm or 3 mm maps . table [ sourcetable ] gives the positions , sizes , and 3 mm fluxes associated with the 30 robust sources . the median value of the emissivity spectral index between 1 mm and 3 mm in omc-2/3 is @xmath46 , with a standard deviation of 0.3 . the typical uncertainty due to noise in the spectral index at any given point in our map of @xmath1 is 0.16 . the systematic uncertainty in @xmath1 due to absolute flux calibration uncertainties is 0.17 . the spectral energy distribution in omc-2/3 is much shallower than has been measured in omc-1 , where @xmath15 @xcite . in the following , we discuss possible explanations for the low values of @xmath1 determined in omc-2/3 . the spectral indices ranged from @xmath47 to @xmath48 ( note that a blackbody would have @xmath49 ) . this variation is significantly larger than the uncertainties in @xmath1 , so the ratio of the 1 mm flux to 3 mm flux is strongly dependent on position . we find no significant differences between the mean values of @xmath1 found towards starless cores , protostellar cores , or at positions within the omc-2/3 filament not associated with a core , with core positions taken from @xcite . the position of possible protostars in omc-2/3 ( asterisks and diamonds in figure [ tempbetafig ] ) , correspond to regions with the full range of steep to shallow seds . there are relatively few points associated with starless cores and filamentary material removed from dense cores , but these regions also have the same average spectral index and spread ( @xmath50 ) . one starless core in particular , j053530.0 - 045850 in the nomenclature of @xcite , is notable for its low emissivity spectral index ( @xmath51 ) . we suggest that this core would be an interesting target for future observations . values of emissivity spectral index with @xmath52 have been seen before in the disks around pre - main sequence stars and brown dwarfs ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) and are attributed to dust grain sizes up to millimeter scales . seven dense cores in the orion b9 region have had their spectral indices measured between 350 @xmath8 850 , with a median value of @xmath53 @xcite . large grains provide a possible explanation for the relatively flat spectral index found in the orion cores in the @xcite sample . the presence of millimeter - sized grains is also given as the likely explanation for @xmath52 found in three class 0 protostellar cores in the 1.3 mm - 2.7 mm wavelength range by @xcite . similarly , the low values of @xmath1 in omc-2/3 can be attributed to much larger dust grains than are typically found in the diffuse ism . the mustang and mambo maps do not recover large - scale features associated with diffuse emission in omc-2/3 , so we have only measured the emissivity spectral index of the more compact ( i.e. , likely the densest ) portions of the filaments as well as the dense starless and protostellar cores . note that @xcite find a typical density for this filament of @xmath610@xmath54@xmath28 , much lower than is found in circumstellar disks but very similar to the densities found by @xcite in cores in orion b9 . in omc-1 , @xcite find that the spectral index decreases with density , which is qualitatively consistent with the idea that the observations presented here of the densest portions of omc-2/3 are characterized by a low average value of @xmath1 . adopting the dust grain model presented in @xcite , we find that a spectral index of @xmath55 indicates that the maximum grain size is at least 1 mm and could be as large is 1 cm depending on the powerlaw slope of the dust grain size distribution . the dust opacity at 1 mm implied by the range of dust grain distributions and maximum grain size is @xmath56 @xmath40g@xmath24 @xcite . this range of @xmath57 values is roughly a factor of 3 - 7 larger than the standard value given @xcite table 1 column 6 . core masses can be determined from the mustang 3 mm map using @xmath58 where d is the distance to omc-2/3 . since the mass derived from equation [ masseq ] varies inversely with the opacity , core masses estimated from their 1 mm flux would be a factor of 3 - 7 lower assuming large grains than assuming standard @xcite grains . taking the median dust emissivity spectral index of @xmath59 , we find that core masses derived from 3 mm emission are a factor of 2.8 smaller than the masses determined with the commonly assumed value of @xmath60 . the mass derived from 3 mm emission , combining the effects of a higher opacity and shallower emissivity spectral index , would be a factor of 10 - 20 lower for a model with large grains and @xmath1 = 0.9 than one would derive from an extrapolation of the @xcite value at 1 mm using and emissivity spectral index of 2 . for example , a typical core in omc-2/3 has an integrated flux of 20mjy ( see table [ sourcetable ] ) at 3 mm . the mass derived for this typical core , assuming a temperature of 16.5k ( see section [ ammonia ] ) , would be 2.8 for @xmath60 and an opacity at 1 mm taken from @xcite . on the other hand , the mass derived from the opacity at 1 mm given in @xcite extrapolated to 3 mm using @xmath46 would be no greater than 0.3 . in our determination of @xmath1 ( see section [ betamap ] ) , we assumed that there is no free - free emission contributing significantly to the observed 1 mm or 3 mm maps . in regions of high - mass star formation , however , the 3 mm ( and even 1 mm ) continuum can have a significant contribution from free - free emission . for instance , towards orion bn / kl in omc-1 ( to the immediate south of omc-2/3 ) , the 3 mm continuum is evenly split between free - free and thermal dust emission and in some places the continuum is actually dominated by free - free emission @xcite . the omc-2/3 region lacks the cluster of ob stars that dominate omc-1 , so we expect the 3 mm emission from omc-2/3 to originate predominantly from dust rather than ionized gas . the expectation that the 3 mm emission is overwhelmingly from dust is borne out by the similar morphology of the mustang and spatially filtered mambo maps ( see figure [ fluxfig ] ) . the assumption that the 3 mm continuum emission from omc-2/3 is purely from dust emission has been made in previous studies ( e.g. , * ? ? ? * ) and this assumption was also made in the recent mustang survey of six class 0 protostars @xcite . @xcite used the vla to map the small - scale 3.6 cm emission toward omc-2/3 , finding 14 sources with flux densities ranging from 0.15mjy to 2.84mjy at an angular resolution of 8 . for optically thin free - free emission ( @xmath61 ) , the contribution of any such emission to the mustang map would be insignificant , given the @xmath60.4mjybeam@xmath24 noise in the map . for optically thick emission , however , @xmath62 and free - free emission becomes significant . we expect some localized emission at the positions of the omc-2/3 protostars , which would bias our derived values of the emissivity spectral index to low values at these locations . the free - free emission measured by @xcite was unresolved in 11 of the 14 detected sources and less than 17 in all cases , so any free - free emission from these sources would be confined to regions one or two resolution elements wide in our maps . we suggest that the emissivity spectral index in regions with @xmath63 is best explained by contamination from free - free emission toward these objects . for example , the few pixels in figure [ tempbetafig ] with @xmath63 generally correspond to known protostellar cores or the 3.6 cm continuum sources ( as marked in figure [ fluxfig ] ) . nevertheless , we do not expect free - free emission on the spatial scales of filaments to be significant , and we see low values of @xmath1 associated with such structures as well . the starless cores , protostellar cores , and the filament all have the same average value of @xmath1 . future observations of wavelengths between 3.3 mm and 3.6 cm would allow us to determine better the free - free contribution to the 3 mm map . in addition to thermal dust emission and free - free emission , anomalous microwave emission ( ame ) can be another important component of the spectral energy distribution . a strong candidate for the source of ame is small , spinning dust grains @xcite . the ame contribution to the total sed of molecular clouds has been estimated in a series of recent papers by the planck collaboration . in the gould belt molecular clouds , the ame typically peaks at [email protected] , falls below thermal dust emission at frequencies higher than @xmath630 - 40ghz , and is orders of magnitude smaller than thermal dust emission at frequencies greater than 80ghz , where mustang is sensitive @xcite . in the perseus molecular cloud , thermal dust emission dominates ame at frequencies greater than 60ghz and the ame contribution at 80ghz and above is statistically insignificant @xcite . similar results are found in the milky way galactic plane @xcite , and a wide range of galactic clouds @xcite . to estimate the contribution of ame to the observed 90ghz emission , we use planck data for omc-2/3 acquired from the nasa ipac infrared science archive . the planck resolution at 30ghz ( @xmath633 ) is large enough to blend emission from omc-1 , which at long wavelengths is dominated by free - free emission , with the emission from omc-2 . we therefore restrict our analysis to planck data between 30ghz and 353ghz in the omc-3 region . we find ( see figure [ planckfig ] ) that the emission falls by three orders of magnitude from 30ghz ( 1 cm ) to 70ghz ( 4 mm ) . therefore , emission at 100ghz ( 3 mm ) and higher frequencies should be dominated by thermal emission from large dust grains . we caution that the angular scales of the planck maps ( @xmath65 10 at frequencies @xmath66 100ghz ) are much larger than the @xmath610 angular scales probed by our mustang and mambo observations . nevertheless , we believe that there is not much contamination from ame in the mustang 3 mm map . higher angular resolution maps of 10 - 70ghz continuum emission are needed , however , to confirm this expectation . one final possible source of contamination in the 1 mm and 3 mm bands is molecular line emission . the main source of molecular line contamination of dust emission is from co , the brightest molecular line in regions like omc-2/3 . for instance , @xcite found that @xmath67co ( 3 - 2 ) emission contributes @xmath6620% of the 850@xmath68 m flux observed with scuba-2 towards molecular clouds in most regions , but can be the dominant source of emission towards regions with molecular outflows . @xmath67co ( 1 - 0 ) , at 115ghz , falls outside the range of frequency range ( 81ghz - 99ghz ) to which mustang is sensitive @xcite . @xmath67co ( 2 - 1 ) , at 230ghz , falls within the mambo bandpass of 210ghz - 290ghz @xcite . given the temperature and density distributions in omc-2/3 , we would expect an integrated intensity of co ( 2 - 1 ) of @xmath66 30kkms@xmath24 @xcite . this would contribute less than 2mjybeam@xmath24 to our 1 mm mambo map of omc-2/3 , which is much less than the noise in the map . an especially bright and broad ( in velocity space ) outflow could possibly make a significant contribution to the 1 mm map , artificially steepening the derived value of @xmath1 in its vicinity . over the majority of the omc-2/3 filament , we do not expect molecular line contamination to significantly affect either the 3 mm or 1 mm continuum measurements of omc-2/3 . in the rayleigh - jeans approximation ( @xmath69 ) , the observed flux at a given frequency is linearly dependent on temperature , so the ratio of fluxes at different frequencies is independent of temperature . the rayleigh - jeans approximation is only loosely valid for our observations of the omc-2/3 region , with @xmath70 values of 12k and 4k at 250ghz and 90ghz , respectively . @xcite found that the temperatures of cores within omc-2/3 are roughly 17k , with evidence for warmer ( @xmath625k ) dust in the filament . the gas kinetic temperature of the omc-2/3 filament was measured using gbt and vla observations of the nh@xmath0 ( 1,1 ) and ( 2,2 ) inversion transitions @xcite . the gas temperatures are found to be mostly in the range of 10k @xmath66 @xmath25 @xmath66 20k . we therefore consider the possibility that line - of - sight temperature variations influence the values of @xmath1 determined from the cold dust in omc-2/3 . when determining the emissivity spectral index map shown in figure [ tempbetafig ] , we assumed that the temperature along each line of sight is constant . the true temperature distribution along each line of sight is very likely to be variable , so here we estimate the importance of line - of - sight temperature variations on our @xmath1 map . we construct a toy model where the dust temperature along a line of sight is given by a gaussian distribution with a 1@xmath2 width of 25% of the mean temperature . we let the mean temperature vary from 5k to 25k in steps of 0.1k . for each distribution , we calculate the ratio of the continuum emission at 1.1 mm and 3 mm , assuming an emissivity spectral index of 1 . the 1mm/3 mm flux ratio depends on the temperature distribution of the dust , especially for low temperatures where the rayleigh - jeans approximation is least valid . we then use equation [ betaeq ] to determine the emissivity spectral index that would be derived from the 1mm/3 mm flux ratio . we find that temperature variations lead to a small overestimate of the emissivity spectral index , i.e. , temperature variations along the line - of - sight push @xmath1 to larger absolute values . the error is 2% ( or less ) for temperature distributions centered on 10k or warmer , and increases to 8% at a mean temperature of 5k ( see figure [ correlation ] ) . we therefore find that line - of - sight temperature variations could systematically bias our derived values of @xmath1 , but at a level much lower than the uncertainties introduced by noise and absolute flux calibration errors . the systematic bias due to line - of - sight temperature variations moves @xmath1 in the opposite direction of our surprising result - i.e. , we find that the sed is shallower than predicted for dust in a filament while line - of - sight temperature variations work to steepen the sed . figure [ correlation ] compares @xmath1 and temperature for the entire omc-2/3 region . we find that @xmath1 and @xmath35 are anti - correlated at the 4@xmath2 level . an anti - correlation between temperature and the emissivity spectral index has been seen previously in some molecular clouds ( e.g. , * ? ? ? * ; * ? ? ? * ) , but other studies have not found conclusive evidence for a temperature - dependent emissivity spectral index ( e.g. , * ? ? ? when using dust emission maps to determine both the temperature and emissivity spectral index , line - of - sight temperature variations can create spurious @xmath35-@xmath1 correlations ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , though these can be at least partially overcome by careful analysis and statistical modeling ( e.g. , * ? ? ? * ; * ? ? ? the analysis here is different from most previous studies of correlations between temperature and @xmath1 in that these two quantities are determined from different data sets ( nh@xmath0 observations and millimeter - wavelength dust emission ) . as shown in figure [ correlation ] , this anti - correlation , if real , is not caused by line - of - sight temperature variations . it is possible that our assumption of a power - law dust emissivity ( i.e. , @xmath71 ) between 90ghz and 250ghz is not justified . for certain size distributions and compositions , detailed models of the dust opacity between 90ghz and 250ghz show that it has a more complicated form ( e.g. , * ? ? ? laboratory measurements and theoretical models of dust grain analogues have found that the emissivity can be temperature - dependent ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) and not well described by a simple power - law ( e.g. , * ? ? ? * ; * ? ? ? the relatively low values of @xmath1 we find might not be a result of large dust grains , but rather may be determined by the grain composition , temperature , and the wavelengths of our observations . observations of circumstellar disks , however , often sample the sed from sub - millimeter to centimeter wavelengths and these are successfully modeled with a simple power - law opacity ( e.g. , * ? ? ? * ; * ? ? ? * ) . a trivial explanation for the low values of @xmath1 found in omc-2/3 would calibration errors . given the relatively large frequency range between the mambo and mustang observations , it would require a large error in the flux calibration to bring the emissivity spectral index into agreement with values found in perseus b1 or the orion nebula cluster . for instance , if we decrease the fluxes in the 3 mm map by 10% and increase the fluxes in the 1 mm map by 10% , then the median value of @xmath1 increases from 0.9 to 1.0 ( the standard deviation of derived values of @xmath1 remains 0.3 ) . to bring the median value of @xmath1 up to 2 , we would need to increase the ratio of @xmath72 by a factor of 4 . we do not believe that the flux uncertainties in the mambo or mustang maps are anywhere near that high , so the absolute calibration of the bolometer maps is not likely to be the main driver of the low values of @xmath1 found here . of the possible explanations for the shallow emissivity spectral index in omc-2/3 , we find that the two most likely are the presence of large grains or a dust emissivity not well described by a single power - law . if dust grains in omc-2/3 are characterized by millimeter size scales , this would be the first report of such large grains in structures with scales on the order of 1pc . the protostellar cores and circumstellar disks with millimeter - sized grains have sizes of @xmath73 0.1pc . it could be important to models of grain growth in disks and cores if the dust in filaments is already quite large . in this case , it will be important to determine if omc-2/3 is unique in exhibiting large grains or if this is a common feature of star - forming filaments . although omc-2/3 is unique in that it has a higher density of starless and protostellar cores than other regions within @xmath6500pc of the sun , there are no other properties ( mass , density , temperature , etc . ) that would lead one to suspect that the dust grains in omc-2/3 ought to have properties significantly different than those found in other nearby molecular clouds . alternatively , the shallow spectral index between 1 mm and 3 mm found in omc-2/3 is due to an sed that is not well - characterized by a power law function at millimeter wavelengths . in this case , the grains are not necessarily larger than is commonly found in nearby molecular clouds . it would be important to take the true shape of the dust emissivity into account when conducting future studies of @xmath1 in filaments . in either case , more detailed observations are needed to study the full dust sed in filaments to make accurate measurements of mass , temperature , and possible large grains . some of the assumptions made in section [ analysis ] and some of the potential sources of error given in section [ discussion ] can be tested with additional observations . we have made the assumption that all of the 3 mm flux comes from dust emission , but it seems clear that free - free emission may contribute somewhat to the observed 3 mm map , especially near protostars . the contribution of free - free emission to the 3 mm map can be tested by making observations at wavelengths between the 3.6 cm map produced by @xcite and the 3.3 mm map shown here . these observations could be used to determine at what wavelength the free - free emission observed by @xcite switches from being optically thick to thin , allowing us to subtract accurately the free - free component from the 3 mm map and make a more accurate map of the emissivity spectral index . we have made the assumption that the dust opacity is given by a power law , but this assumption can be tested by additional observations . to avoid uncertainties introduced by line - of - sight temperature variations , observations at 850@xmath68 m ( with scuba-2 , for example ) or 2 mm ( with the iram 30 m , for example ) could be used to constrain better the shape of the ( sub)millimeter sed and minimize the effect of uncertainties in absolute flux calibration on the derived value of the emissivity spectral index . herschel observations between 70 and 500 could also be folded into this analysis , with similar spatial filtering and careful corrections for temperature variations and optical depth . high - resolution extinction maps could also be used to constrain the dust opacity independently from the dust temperature distribution . we have mapped the emissivity spectral index of the omc-2/3 using new 3 mm observations paired with previously published observations of the 1 mm continuum and nh@xmath0-derived gas temperature . focusing more on spatial scales of filaments and cores , we find that the emissivity spectral index is much shallower than is often assumed for dust in molecular clouds , with @xmath74 . we find a weak correlation between @xmath1 and @xmath35 , though there are no significant differences between the average values of @xmath1 found in the omc-2/3 filament , in the starless cores , and in the protostellar cores . such a low average value of @xmath1 in omc-2/3 could be explained by millimeter - sized dust grains , as inferred in many circumstellar disks . dust emissivities that vary with temperature or opacities that do not vary with wavelength as a simple power law could also explain our observations . we do not expect either free - free emission or ame to be significant on the scale of the 2 parsec - long omc-2/3 filament , though , if present and sufficiently bright , either emission mechanism would result in lower absolute values of @xmath1 . we suggest future observations that can be used to determine the strength of free - free emission and ame . we thank our anonymous referee for comments that improved the content and clarity of this paper . the national radio astronomy observatory is a facility of the national science foundation operated under cooperative agreement by associated universities , inc . jdf acknowledges support by the national research council of canada and the natural sciences and engineering council of canada(nserc ) via a discovery grant . rf is a dunlap fellow at the dunlap institute for astronomy and astrophysics , university of toronto . the dunlap institute is funded through an endowment established by the david dunlap family and the university of toronto . dl is supported by china ministry of science and technology under state key development program for basic research ( 2012cb821800 ) . the authors would like to thank the mustang instrument team from the university of pennsylvania , nrao , cardiff university , nasa - gsfc , and nist for their efforts on the instrument and software that have made this work possible . d. li acknowledges the support from national basic research program of china ( 973 program ) no . 2012cb821800 , nsfc no . 11373038 , and chinese academy of sciences grant no . xdb09000000 . agladze , n. i. , sievers , a. j. , jones , s. a. , burlitch , j. m. , & beckwith , s. v. w. 1996 , , 462 , 1026 arab , h. , abergel , a. , habart , e. , et al . 2012 , , 541 , a19 bertoldi , f. , carilli , c. , aravena , m. , et al . 2007 , , 172 , 132 boudet , n. , mutschke , h. , nayral , c. , et al . 2005 , , 633 , 272 coupeaud , a. , demyk , k. , meny , c. , et al . 2011 , , 535 , a124 davis , c. j. , froebrich , d. , stanke , t. , et al . 2009 , , 496 , 153 di francesco , j. , johnstone , d. , kirk , h. , mackenzie , t. , & ledwosinska , e. 2008 , , 175 , 277 dicker , s. r. , korngut , p. m. , mason , b. s. , et al . 2008 , , 7020 , dicker , s. r. , mason , b. s. , korngut , p. m. , et al . 2009 , , 705 , 226 doty , s. d. , & neufeld , d. a. 1997 , , 489 , 122 drabek , e. , hatchell , j. , friberg , p. , et al . 2012 , , 426 , 23 draine , b. t. , & lazarian , a. 1998 , , 508 , 157 draine , b. t. 2006 , , 636 , 1114 draine , b. t. , & hensley , b. 2012 , , 757 , 103 dupac , x. , bernard , j .- p . , boudet , n. , et al . 2003 , , 404 , l11 friesen , r. k. , johnstone , d. , naylor , d. a. , & davis , g. r. 2005 , , 361 , 460 goldsmith , p. f. , bergin , e. a. , & lis , d. c. 1997 , , 491 , 615 goldsmith , p. f. 2001 , , 557 , 736 greve , a. , kramer , c. , & wild , w. 1998 , , 133 , 271 ho , p. t. p. , & townes , c. h. 1983 , , 21 , 239 hunter , t. r. 1998 , , 110 , 634 johnstone , d. , & bally , j. 1999 , , 510 , l49 juvela , m. , & ysard , n. 2012a , , 539 , a71 juvela , m. , & ysard , n. 2012b , , 541 , a33 juvela , m. , montillaud , j. , ysard , n. , & lunttila , t. 2013 , , 556 , a63 kelly , b. c. , shetty , r. , stutz , a. m. , et al . 2012 , , 752 , 55 kim , m. k. , hirota , t. , honma , m. , et al . 2008 , , 60 , 991 kwon , w. , looney , l. w. , mundy , l. g. , chiang , h .- f . , & kemball , a. j. 2009 , , 696 , 841 li , d. , goldsmith , p. f. , & menten , k. 2003 , , 587 , 262 li , d. , kauffmann , j. , zhang , q. , & chen , w. 2013 , , 768 , l5 lis , d. c. , serabyn , e. , keene , j. , et al . 1998 , , 509 , 299 lommen , d. , wright , c. m. , maddison , s. t. , et al . 2007 , , 462 , 211 malinen , j. , juvela , m. , collins , d. c. , lunttila , t. , & padoan , p. 2011 , , 530 , a101 mason , b. s. , dicker , s. r. , korngut , p. m. , et al . 2010 , , 716 , 739 mennella , v. , brucato , j. r. , colangeli , l. , et al . 1998 , , 496 , 1058 menshchikov , a. , andr , p. , didelon , p. , et al . 2012 , , 542 , a81 menten , k. m. , reid , m. j. , forbrich , j. , & brunthaler , a. 2007 , , 474 , 515 meny , c. , gromov , v. , boudet , n. , et al . 2007 , , 468 , 171 miettinen , o. , harju , j. , haikala , l. k. , & juvela , m. 2012 , , 538 , a137 nutter , d. , & ward - thompson , d. 2007 , , 374 , 1413 ossenkopf , v. , & henning , t. 1994 , , 291 , 943 paradis , d. , bernard , j .- p . , mny , c. , & gromov , v. 2011 , , 534 , a118 prez , l. m. , carpenter , j. m. , chandler , c. j. , et al . 2012 , , 760 , l17 peterson , d. e. 2005 , ph.d . thesis planck collaboration , ade , p. a. r. , aghanim , n. , et al . 2011 , , 536 , a20 planck collaboration , abergel , a. , ade , p. a. r. , et al . 2011 , , 536 , a21 planck collaboration , abergel , a. , ade , p. a. r. , et al . 2011 , , 536 , a25 planck collaboration , ade , p. a. r. , aghanim , n. , et al . 2013a , arxiv:1301.5839 planck collaboration , ade , p. a. r. , aghanim , n. , et al . 2013b , arxiv:1309.1357 reipurth , b. , rodrguez , l. f. , & chini , r. 1999 , , 118 , 983 ricci , l. , testi , l. , natta , a. , et al . 2010 , , 512 , a15 ricci , l. , testi , l. , natta , a. , scholz , a. , & de gregorio - monsalvo , i. 2012 , , 761 , l20 sadavoy , s. i. , di francesco , j. , bontemps , s. , et al . 2010 , , 710 , 1247 sadavoy , s. i. , di francesco , j. , johnstone , d. , et al . 2013 , arxiv:1303.1529 schnee , s. , enoch , m. , noriega - crespo , a. , et al . 2010 , , 708 , 127 shetty , r. , kauffmann , j. , schnee , s. , goodman , a. a. , & ercolano , b. 2009a , , 696 , 2234 shetty , r. , kauffmann , j. , schnee , s. , & goodman , a. a. 2009b , , 696 , 676 shimajiri , y. , takahashi , s. , takakuwa , s. , saito , m. , & kawabe , r. 2009 , , 61 , 1055 shirley , y. l. , nordhaus , m. k. , grcevich , j. m. , et al . 2005 , , 632 , 982 shirley , y. l. , mason , b. s. , mangum , j. g. , et al . 2011 , , 141 , 39 tabatabaei , f. s. , braine , j. , xilouris , e. m. , et al . 2013 , arxiv:1310.4155 ubach , c. , maddison , s. t. , wright , c. m. , et al . 2012 , , 425 , 3137 van der tak , f. f. s. , black , j. h. , schier , f. l. , jansen , d. j. , & van dishoeck , e. f. 2007 , , 468 , 627 veneziani , m. , piacentini , f. , noriega - crespo , a. , et al . 2013 , , 772 , 56 weiland , j. l. , odegard , n. , hill , r. s. , et al . 2011 , , 192 , 19 williams , j. p. , de geus , e. j. , & blitz , l. 1994 , , 428 , 693 yang , m. , & phillips , t. 2007 , , 662 , 284 ysard , n. , juvela , m. , demyk , k. , et al . 2012 , , 542 , a21 ccccccc 83.8473 & -5.0253 & 10.8 & 10.8 & 161.9 & 0.112 & 0.146 + 83.8613 & -5.1662 & 20.8 & 14.0 & 35.2 & 0.036 & 0.109 + 83.8650 & -5.1587 & 18.0 & 10.8 & 157.6 & 0.028 & 0.053 + 83.8522 & -5.1311 & 16.0 & 10.8 & 96.9 & 0.026 & 0.050 + 83.8584 & -5.0952 & 10.8 & 10.8 & 140.3 & 0.020 & 0.027 + 83.8599 & -5.0656 & 10.8 & 10.8 & 28.4 & 0.018 & 0.024 + 83.8427 & -5.0202 & 12.8 & 10.8 & 129.8 & 0.020 & 0.033 + 83.8513 & -5.1419 & 11.4 & 10.8 & 149.2 & 0.018 & 0.026 + 83.8250 & -5.0054 & 13.1 & 10.8 & 134.5 & 0.014 & 0.025 + 83.8472 & -5.2006 & 10.8 & 10.8 & 103.5 & 0.017 & 0.020 + 83.8338 & -5.2204 & 10.8 & 10.8 & 110.5 & 0.018 & 0.020 + 83.8736 & -4.9800 & 10.8 & 10.8 & 100.9 & 0.018 & 0.029 + 83.8397 & -5.2206 & 14.9 & 10.8 & 48.4 & 0.016 & 0.030 + 83.8259 & -5.0096 & 10.8 & 10.8 & 178.1 & 0.015 & 0.018 + 83.8598 & -5.1724 & 13.2 & 10.8 & 179.5 & 0.015 & 0.022 + 83.8353 & -5.0144 & 21.1 & 10.9 & 120.9 & 0.007 & 0.017 + 83.8449 & -5.2071 & 10.8 & 10.8 & 160.0 & 0.012 & 0.015 + 83.8480 & -5.1186 & 13.3 & 11.3 & 100.1 & 0.010 & 0.019 + 83.8456 & -5.2110 & 27.4 & 10.8 & 36.5 & 0.009 & 0.022 + 83.8605 & -5.0875 & 18.8 & 10.8 & 15.8 & 0.007 & 0.015 + 83.8530 & -5.1745 & 10.8 & 10.8 & 105.4 & 0.006 & 0.006 + 83.8736 & -4.9966 & 10.8 & 10.8 & 173.4 & 0.008 & 0.010 + 83.8294 & -5.0145 & 15.3 & 10.8 & 114.2 & 0.006 & 0.012 + 83.8664 & -5.1180 & 13.9 & 11.3 & 122.7 & 0.008 & 0.018 + 83.8670 & -5.1747 & 14.4 & 10.8 & 149.4 & 0.006 & 0.009 + 83.8469 & -5.1253 & 16.8 & 12.3 & 148.5 & 0.004 & 0.009 + 83.8619 & -5.1022 & 21.1 & 10.8 & 152.8 & 0.005 & 0.008 + 83.8156 & -4.9987 & 31.1 & 12.3 & 121.8 & 0.003 & 0.010 + 83.8552 & -5.0428 & 24.7 & 13.3 & 154.0 & 0.004 & 0.012 + 83.8333 & -5.0847 & 21.9 & 11.1 & 114.4 & 0.003 & 0.007 +	we present a new 3.3 mm continuum map of the omc-2/3 region . when paired with previously published maps of 1.2 mm continuum and nh@xmath0-derived temperature , we derive the emissivity spectral index of dust emission in this region , tracking its changes across the filament and cores . we find that the median value of the emissivity spectral index is 0.9 , much shallower than previous estimates in other nearby molecular clouds . we find no significant difference between the emissivity spectral index of dust in the omc-2/3 filament and the starless or protostellar cores . furthermore , the temperature and emissivity spectral index , @xmath1 , are anti - correlated at the 4@xmath2 level . the low values of the emissivity spectral index found in omc-2/3 can be explained by the presence of millimeter - sized dust grains in the dense regions of the filaments to which these maps are most sensitive . alternatively , a shallow dust emissivity spectral index may indicate non - powerlaw spectral energy distributions , significant free - free emission , or anomalous microwave emission . we discuss the possible implications of millimeter - sized dust grains compared to the alternatives . ism : dust ism : clouds stars : formation stars : protostars
the problem of convergence of discrete - time financial models to the models with continuous time is well developed ; see , e.g. , @xcite . the reason for such an interest can be explained as follows : from the analytical point of view , it is much simpler to deal with continuous - time models although all real - world models operate in the discrete time . in what concerns the rate of convergence , there can be different approaches to its estimation . some of this approaches are established in @xcite . in this paper , we consider the cox ingersoll ross process and its approximation on a finite time interval . the cir process was originally proposed by cox , ingersoll , and ross @xcite as a model for short - term interest rates . nowadays , this model is widely used in financial modeling , for example , as the volatility process in the heston model @xcite . the strong global approximation of cir process is studied in several articles . strong convergence ( without a rate or with a logarithmic rate ) of several discretization schemes is shown by @xcite . in @xcite , a general framework for the analysis of strong approximation of the cir process is presented along with extensive simulation studies . nonlogarithmic convergence rates are obtained in @xcite . in @xcite , the author extends the cir model of the short interest rate by assuming a stochastic reversion level , which better reflects the time dependence caused by the cyclical nature of the economy or by expectations concerning the future impact of monetary policies . in this framework , the convergence of the long - term return by using the theory of generalized bessel - square processes is studied . in @xcite , the authors propose an empirical method that utilizes the conditional density of the state variables to estimate and test a term structure model with known price formula using data on both discount and coupon bonds . the method is applied to an extension of a two - factor model due to cox , ingersoll , and ross . their results show that estimates based solely on bills imply unreasonably large price errors for longer maturities . the process is also discussed in @xcite . in this article , we focus on the regime where the cir process does not hit zero and study weak approximation of this process . in the first case , the sequence of prelimit markets is modeled as the sequence of the discrete - time additive stochastic processes , whereas in the second case , the sequence of multiplicative stochastic processes is modeled . the additive scheme is widely used , for example , in the papers @xcite . the papers @xcite are recent examples of modeling a stochastic interest rate by the multiplicative model of cir process . in @xcite , the authors say that the model has the `` strong convergence property , '' whereas they refer to models as having the `` weak convergence property '' when the returns converge to a constant , which generally depends upon the current economic environment and that may change in a stochastic fashion over time . we construct a discrete approximation scheme for the price of asset that is modeled by the cox ingersoll ross process . in order to construct these additive and multiplicative processes , we take the euler approximations of the cir process itself but replace the increments of the wiener process with iid bounded vanishing symmetric random variables . we introduce a `` truncated '' cir process and use it to prove the weak convergence of asset prices . the paper is organized as follows . in section [ section2 ] , we present a complete and `` truncated '' cir process and establish that the `` truncated '' cir process can be described as the unique strong solution to the corresponding stochastic differential equation . we establish that this `` truncated '' process does not hit zero under the same condition as for the original nontruncated process . in section [ section3 ] , we present discrete approximation schemes for both these processes and prove the weak convergence of asset prices for the additive model . in the next section , we prove the weak convergence of asset prices for the multiplicative model . appendix contains additional and technical results.=1 let @xmath0 be a complete filtered probability space , and @xmath1 be an adapted wiener process . consider a cox ross process with constant parameters on this space . this process is described as the unique strong solution of the following stochastic differential equation : @xmath2 where @xmath3 , @xmath4 . the integral form of the process @xmath5 has the following form : @xmath6 according to the paper @xcite , the condition @xmath7 is necessary and sufficient for the process @xmath8 to get positive values and not to hit zero . further , we will assume that this condition is satisfied . for the proof of functional limit theorems , we will need a modification of the cox ingersoll ross process with bounded coefficients . this process is called a truncated cox ingerssol ross process . let @xmath9 . consider the following stochastic differential equation with the same coefficients @xmath10 and @xmath11 : @xmath12 [ lem1 ] for any has a unique strong solution . since the coefficients @xmath13 and @xmath14 satisfy the conditions of theorem [ watanabe2 ] and also the growth condition , a global strong solution @xmath15 exists uniquely for every given initial value @xmath16 . denote @xmath17 with @xmath18 such that @xmath19 . suppose that @xmath20 . then for any @xmath21 such that @xmath22 for @xmath23 , we would have , with positive probability , @xmath24 on the interval @xmath25 , and hence @xmath26 would increase in this interval . this is obviously impossible . therefore , @xmath15 is nonnegative and can be written as @xmath27 the integral form of the process @xmath28 is as follows : @xmath29 [ lem2 ] let @xmath30 and @xmath31 . then the trajectories of the process @xmath28 are positive with probability 1 . in order to prove that the process @xmath28 is positive , we will use the proof similar to that given in @xcite for the complete cox ingersoll ross process with corresponding modifications . note that the coefficients @xmath32 and @xmath33 are continuous and @xmath34 on @xmath35 . fix @xmath36 and @xmath37 such that @xmath38 . due to the nonsingularity of @xmath39 on @xmath40 $ ] , there exists a unique solution @xmath41 of the ordinary differential equation @xmath42 with boundary conditions @xmath43 , and this solution is nonnegative , which follows from its representation through a nonnegative green function given in @xcite . define the stopping times @xmath44 by the it formula , for any @xmath45 , @xmath46 this formula and nonnegativity of @xmath47 imply that @xmath48 and , as @xmath49 , @xmath50 this means that @xmath28 exits from every compact subinterval of @xmath51\subset(0,\infty)$ ] in finite time . it follows from the boundary conditions and equality @xmath52 that @xmath53 , and then from we have @xmath54 let us now define the function @xmath55 which has a continuous strictly positive derivative @xmath56 , and the second derivative @xmath57 exists and satisfies @xmath58 . the it formula shows that , for any @xmath45 , @xmath59 and @xmath60 taking the limit as @xmath49 , we get @xmath61 and hence @xmath62 consider the integral @xmath63 first , consider the case @xmath64 . then @xmath65 and if @xmath7 , then @xmath66 now let @xmath67 increase and tend to infinity . denote @xmath68 . then , for @xmath69,@xmath70 and thus @xmath71 . define @xmath72 and put @xmath73 . from we get @xmath74 and , as @xmath75 , we get that , for any @xmath76 , @xmath77 , whence , finally , @xmath78 . similarly , @xmath79 . assume now that @xmath80 . then @xmath81 so the events @xmath82 and @xmath83 can not both have probability 1 . this contradiction shows that @xmath84 , whence @xmath85 if @xmath30 . now , let @xmath86 be fixed . @xmath87 : x_t\neq x_t^c \bigr\}\rightarrow0\ ] ] as @xmath88 . obviously , it suffices to show that @xmath89}\|x_t\|\ge c \bigr\}\rightarrow0 \quad \text{as } \ ; c\rightarrow\infty.\ ] ] it is well known ( see , e.g. , @xcite ) that @xmath90 follows a noncentral @xmath91 distribution with ( in general ) noninteger degree of freedom @xmath92 and noncentrality parameter @xmath93 . the first and second moments for any @xmath94 are given by @xmath95 @xmath96therefore , there exists a constant @xmath97 such that @xmath98 , whence @xmath99 , @xmath100 . using the doob inequality , we estimate @xmath101}\|x_t\|\ge c \bigr\}\le\frac{1}{c^2}\operatorname{\mathsf e}\sup\limits_{t\in[0,t]}x_t^2\\ & \quad = \frac{1}{c^2}\operatorname{\mathsf e}\sup\limits_{t\in[0,t ] } \biggl\ { \biggl(x_0+\int\limits_0^t ( b - x_s ) ds+{\sigma}\int\limits_0^t\sqrt{x_s}dw_s \biggr)^2 \biggr\}\\ & \quad \le\frac{3}{c^2 } \biggl\{x_0 ^ 2+t\operatorname{\mathsf e}\biggl(\int\limits_0^t{\vert}b - x_s{\vert}ds \biggr)^2+{\sigma}^2\operatorname{\mathsf e}\sup\limits_{t\in[0,t ] } \biggl(\int\limits_0^t\sqrt { x_s}dw_s \biggr)^2 \biggr\}\\ & \quad \le\frac{3}{c^2 } \biggl\{x_0 ^ 2+t\operatorname{\mathsf e}\int\limits_0^t ( b - x_s ) ^2ds+4{\sigma}^2\operatorname{\mathsf e}\int\limits_0^tx_sds \biggr\}\le\frac{b_1}{c^2}\end{aligned}\ ] ] for some constant @xmath102 . the lemma is proved . consider the following discrete approximation scheme for the process @xmath5 . assume that we have a sequence of the probability spaces @xmath103 , @xmath104 . let @xmath105 , @xmath106 be the sequence of symmetric iid random variables defined on the corresponding probability space and taking values @xmath107 , that is , @xmath108 . let further @xmath109 . we construct discrete approximation schemes for the stochastic processes @xmath5 and @xmath28 as follows . consider the following approximation for the complete process : @xmath110 and the corresponding approximations for @xmath28 given by @xmath111 the following lemma confirms the correctness of the construction of these approximations . let , @xmath112 . * if @xmath113 , then all values given and are positive . * we have @xmath114 as @xmath88 . @xmath115 we apply the method of mathematical induction . when @xmath116 , @xmath117 let us show that @xmath118 we denote @xmath119 and reduce to the quadratic inequality @xmath120 which obviously holds because the discriminant @xmath121 when @xmath7 and @xmath122 . so , @xmath123 . assume now that @xmath124 . it can be shown by applying the same transformation that when @xmath7 and @xmath122 , the values @xmath125 . it can be proved similarly that the values given by are positive . \2 ) @xmath126 can be represented as @xmath127 compute @xmath128\operatorname{\mathsf e}x_{i-1}^{(n ) } \nonumber\\ & \quad + \biggl(1-\frac{t}{n } \biggr)^2\operatorname{\mathsf e}\bigl(x_{i-1}^{(n ) } \bigr)^2.\label{exkn2}\end{aligned}\ ] ] assume that @xmath129 , @xmath130 , for some @xmath131 . then @xmath132 , @xmath130 . we get the quadratic inequality of the form @xmath133 \beta+ \biggl(\frac{bt}{n } \biggr)^2<\beta^2\ ] ] or , equivalently , @xmath134\beta+ \biggl(\frac { bt}{n } \biggr)^2<0,\ ] ] which obviously holds when @xmath135 . so , for all @xmath136 , @xmath137 . using the burkholder inequality , we estimate @xmath138 therefore , @xmath139 whence the proof follows . consider the sequences of step processes corresponding to these schemes : @xmath140 and @xmath141 thus , the trajectories of the processes @xmath142 and @xmath143 have jumps at the points @xmath144 , and are constant on the interior intervals . consider the filtrations @xmath145 . the processes @xmath143 are adapted with respect to them . therefore , we can consider the same filtrations for all discrete approximation schemes . so , we can identify @xmath146 with @xmath147 for @xmath148 . [ zau3.1 ] now we can rewrite relation as follows : @xmath149 : x_t^{(n)}\neq x_t^{(n , c ) } \bigr\ } \rightarrow0\ ] ] as @xmath88 . denote by @xmath150 and @xmath151 , the measures corresponding to the processes @xmath5 and @xmath152 , respectively , and by @xmath153 and @xmath154 , the measures corresponding to the processes @xmath28 and @xmath155 , respectively . denote by @xmath156 the weak convergence of measures corresponding to stochastic processes . we apply theorem 3.2 from @xcite to prove the weak convergence of measures @xmath157 to the measure @xmath153 . this theorem can be formulated as follows . [ conditions_e ] assume that the following conditions are satisfied : * for any @xmath18 , @xmath158 * for any @xmath18 and @xmath159 $ ] , @xmath160}\operatorname{\mathsf e}\bigl(q_k^{(n , c)}{\mathbb{i}}_{\|q_k^{(n , c)}\|\le a } \big{\vert}{\mathcal{f}}_{k-1}^n \bigr)\\ & \qquad -\int\limits_{0}^{t}\bigl(b - x_s^{(n , c)}\wedge c\bigr)ds\biggr{\vert}\ge \epsilon \biggr)=0;\end{aligned}\ ] ] * for any @xmath18 and @xmath159 $ ] , @xmath161 } \bigl(\operatorname{\mathsf e}\bigl ( \bigl(q_k^{(n , c)}\bigr)^2{\mathbb{i}}_{\|q_k^{(n , c)}\|\le a } \big{\vert}{\mathcal{f}}_{k-1}^n\bigr ) \\ & \qquad - \bigl(\operatorname{\mathsf e}\bigl(q_k^{(n , c)}{\mathbb{i}}_{\|q_k^{(n , c)}\|\le a } \big{\vert}{\mathcal{f}}_{k-1}^n \bigr ) \bigr)^2 \bigr)-{\sigma}^2\int\limits_{0}^{t}\bigl(x_s^{(n , c)}\wedge c\bigr ) ds\biggr{\vert}\ge\epsilon \biggr)=0;\end{aligned}\ ] ] then @xmath162 . using theorem [ conditions_e ] , we prove the following result . [ weak_conv_e ] @xmath162 . according to theorem [ conditions_e ] , we need to check conditions ( i)(iii ) . relation implies that @xmath163 . hence , there exists a constant @xmath164 such that @xmath165 . this means that condition ( i ) is satisfied . furthermore , in order to establish ( ii ) , we consider any fixed @xmath166 and @xmath167 such that @xmath168 , that is , @xmath169 . for such @xmath170 , @xmath171 for any @xmath18 , we have @xmath160}\operatorname{\mathsf e}\bigl(q_k^{(n , c)}{\mathbb{i}}_{\|q_k^{(n , c)}\|\le a } \bigr{\vert}{\mathcal{f}}_{k-1}^n \bigr)-\int\limits_{0}^{t}\bigl(b-\bigl(x_s^{(n , c)}\wedge c\bigr)\bigr)ds\biggr{\vert}\ge\epsilon \biggr)\\ & \quad = \lim\limits_{n}\operatorname{\mathsf p}^n \biggl(\sup\limits_{t\in{\mathbb{t}}}\biggl{\vert}\sum\limits_{1\le k\le [ \frac{nt}{t } ] } \frac{(b-(x_{k-1}^{(n , c)}\wedge c))t}{n}-\sum\limits_{0\le k\le [ \frac{nt}{t}]-1}\bigl(b-\bigl(x_{k}^{(n , c)}\wedge c\bigr)\bigr)\frac{t}{n}\\ & \qquad -\bigl(b-\bigl(x _ { [ \frac{nt}{t } ] } ^{(n , c)}\wedge c\bigr)\bigr ) \biggl(t-\frac { [ \frac{nt}{t } ] t}{n } \biggr)\biggr{\vert}\ge\epsilon \biggr)\\ & \quad = \lim\limits_{n}\operatorname{\mathsf p}^n \biggl(\sup\limits_{t\in{\mathbb{t}}}\biggl{\vert}\bigl(b-\bigl(x _ { [ \frac{nt}{t } ] } ^{(n , c)}\wedge c\bigr)\bigr ) \biggl(t-\frac { [ \frac{nt}{t}]t}{n } \biggr)\biggr{\vert}\ge\epsilon \biggr)=0,\end{aligned}\ ] ] and hence condition ( ii ) is satisfied . now let us check condition ( iii ) . we have @xmath172 therefore , for any @xmath18 , @xmath161 } \bigl(\operatorname{\mathsf e}\bigl ( \bigl(q_k^{(n , c)}\bigr)^2{\mathbb{i}}_{\|q_k^{(n , c)}\|\le a } \bigr{\vert}{\mathcal{f}}_{k-1}^n\bigr ) \\ & \qquad - \bigl(\operatorname{\mathsf e}\bigl(q_k^{(n , c)}{\mathbb{i}}_{\|q_k^{(n , c)}\|\le a } \big{\vert}{\mathcal{f}}_{k-1}^n \bigr ) \bigr)^2 \bigr)-{\sigma}^2\int\limits_{0}^{t}\bigl(x_s^{(n , c)}\wedge c\bigr ) ds\biggr{\vert}\ge\epsilon \biggr)\\ & \quad = \lim\limits_{n}\operatorname{\mathsf p}^n \biggl(\sup\limits_{t\in{\mathbb{t}}}\biggl{\vert}\sum\limits_{1\le k\le [ \frac{nt}{t } ] } \biggl ( \biggl(\frac{(b-(x_{k-1}^{(n , c)}\wedge c ) ) t}{n}\biggr)^2+{\sigma}^2\frac{t}{n } \bigl(x_{k-1}^{(n , c)}\wedge c \bigr)\\ & \qquad - \biggl(\frac { ( b-(x_{k-1}^{(n , c)}\wedge c))t}{n } \biggr)^2 \biggr)-\sum\limits_{0\le k\le [ \frac{nt}{t}]-1 } \biggl({\sigma}^2 \frac{t}{n}\bigl(x_{k}^{(n , c)}\wedge c \bigr ) \biggr)\\ & \qquad -{\sigma}^2 \bigl(x _ { [ \frac{nt}{t } ] } ^{(n , c)}\wedge c \bigr)\biggl(t-\frac { [ \frac{nt}{t } ] t}{n } \biggr)\biggr{\vert}\ge \epsilon \biggr)\\ & \quad = \lim\limits_{n}\operatorname{\mathsf p}^{n } \biggl(\sup\limits_{t\in{\mathbb{t } } } \biggl({\sigma}^2 \bigl(x _ { [ \frac{nt}{t } ] } ^{(n , c)}\wedge c \bigr ) \biggl(t-\frac{[\frac{nt}{t } ] t}{n } \biggr ) \biggr)\ge\epsilon \biggr)=0.\end{aligned}\ ] ] the theorem is proved . [ weak_con ] @xmath173 , @xmath174 . according to theorem [ bill ] and theorem [ weak_conv_e ] , it suffices to prove that @xmath175 however , due to remark [ zau3.1 ] , @xmath176 : x_t^{(n)}\neq x_t^{(n , c ) } \bigr\}=0.\qedhere\end{aligned}\ ] ] in this section , we construct a multiplicative discrete approximation scheme for the process @xmath177 , @xmath178 $ ] , where @xmath179 is the cir process given . we construct the following multiplicative process based on the discrete approximation scheme . we introduce limit and prelimit processes as follows : & s_t^n , c=\{x_0 } _ 1k ( 1+q_k^(n , c ) ) , t , & + & s_t^c= , t , & + & s_t^n=\{x_0 } _ 1k ( 1+q_k^(n ) ) , t , & + & s_t= , t , & + & _ t^n=\{x_0 } _ 1k , t , & + & & + & _ t=\{x_t } , t. & denote by @xmath180 , @xmath181 , @xmath182 , @xmath183 , @xmath184 , and @xmath185 , @xmath104 , the measures corresponding to the processes @xmath186 , @xmath187 , @xmath188 , @xmath189 , @xmath190 , and @xmath191 , @xmath104 , respectively . we apply theorem 3.3 from @xcite to prove the weak convergence of measures . this theorem can be formulated as follows . [ conditions ] let the following conditions hold : * @xmath192 ; * for any @xmath159 $ ] , @xmath193 * for any @xmath159 $ ] , @xmath194 * for any @xmath18 and @xmath159 $ ] , @xmath160}\operatorname{\mathsf e}\bigl(q_k^{(n , c)}{\mathbb{i}}_{\|q_k^{(n , c)}\|\le a } \bigr{\vert}{\mathcal{f}}_{k-1}^n \bigr)\\ & \qquad -\int\limits_{0}^{t}\bigl(b - x_s^{(n , c)}\wedge c\bigr)ds\biggr{\vert}\ge \epsilon \biggr)=0;\end{aligned}\ ] ] * for any @xmath18 and @xmath159 $ ] , @xmath161}\operatorname{\mathsf e}\bigl ( \bigl(q_k^{(n , c)}\bigr)^2{\mathbb{i}}_{\|q_k^{(n , c)}\|\le a } \bigr{\vert}{\mathcal{f}}_{k-1}^n\bigr)\\ & \qquad -{\sigma}^2\int\limits_{0}^{t}\bigl(x_s^{(n , c)}\wedge c\bigr ) ds\biggr{\vert}\ge \epsilon \biggr)=0.\end{aligned}\ ] ] then @xmath195 we prove the following result using theorem [ conditions ] . [ weak_conv ] @xmath196 . according to theorem [ conditions ] , we need to check conditions ( i)(v ) . it was established in the proof of theorem [ weak_conv_e ] that conditions ( i ) and ( iv ) are satisfied . let us show that condition ( ii ) holds . it was also established in the proof of theorem [ weak_conv_e ] that @xmath197 . so , for all @xmath198 $ ] , starting from some number @xmath170 , we have @xmath199 whence condition ( ii ) holds . now , implies that , for all @xmath159 $ ] , starting from some number @xmath170 , we have @xmath200 whence condition ( iii ) holds . let us check condition ( v ) . for any @xmath18 and @xmath201 $ ] , we have @xmath161}\operatorname{\mathsf e}\bigl ( \bigl(q_k^{(n , c)}\bigr)^2{\mathbb{i}}_{\|q_k^{(n , c)}\|\le a } \bigr{\vert}{\mathcal{f}}_{k-1}^n\bigr)-{\sigma}^2\int\limits_{0}^{t}\bigl(x_s^{(n , c)}\wedge c\bigr ) ds\biggr{\vert}\ge\epsilon \biggr)\\ & \quad = \lim\limits_{n}\operatorname{\mathsf p}^n \biggl(\sup\limits_{t\in{\mathbb{t}}}\biggl{\vert}\sum\limits_{1\le k\le [ \frac{nt}{t } ] } \biggl ( \biggl(\frac{(b-(x_{k-1}^{(n , c)}\wedge c ) ) t}{n}\biggr)^2+{\sigma}^2\frac{t}{n } \bigl(x_{k-1}^{(n , c)}\wedge c \bigr ) \biggr)\\ & \qquad - \sum\limits_{0\le k\le [ \frac{nt}{t } ] -1 } \biggl({\sigma}^2 \frac{t}{n}\bigl(x_{k}^{(n , c)}\wedge c \bigr ) \biggr)-{\sigma}^2\bigl(x _ { [ \frac{nt}{t } ] } ^{(n , c)}\wedge c \bigr ) \biggl(t-\frac{[\frac{nt}{t } ] t}{n}\biggr)\biggr{\vert}\ge\epsilon \biggr)\\ & \quad \le\lim\limits_{n}\operatorname{\mathsf p}^{n } \biggl(\sup\limits_{t\in{\mathbb{t } } } \biggl(\frac{(\|b\|+c)^2tt}{n}+{\sigma}^2 \bigl(x _ { [ \frac{nt}{t } ] } ^{(n , c)}\wedge c\bigr ) \biggl(t-\frac { [ \frac{nt}{t } ] t}{n } \biggr ) \biggr)\ge \epsilon \biggr)=0.\end{aligned}\ ] ] the theorem is proved . @xmath202 . the proof immediately follows from theorem [ bill ] , theorem [ weak_conv ] , and remark [ zau3.1 ] . indeed , @xmath176 : x_t^{(n)}\neq x_t^{(n , c ) } \bigr\}=0.\qedhere\end{aligned}\ ] ] the weak convergence @xmath203 can be proved in a similar way . we state here theorem 4.2 from @xcite : [ bill ] suppose that we have sets of processes @xmath204 , @xmath205 , @xmath206 and a stochastic process @xmath5 on the interval @xmath207 $ ] . let @xmath157 , @xmath208 , @xmath209 , and @xmath210 be their corresponding measures . suppose that , for any @xmath9 , @xmath162 , @xmath211 , and that @xmath212 as @xmath88 . suppose further that , for any @xmath18 , @xmath213 then @xmath173 , @xmath211 . [ watanabe1]@xcite . if @xmath11 and @xmath10 are continuous functions satisfying the condition @xmath217 for some positive constant @xmath218 , then for any solution of such that @xmath219 , we have @xmath220 for all @xmath45 . * there exists a strictly increasing function @xmath221 on @xmath222 such that @xmath223 for all @xmath224 . * there exists an increasing and concave function @xmath225 on @xmath222 such that @xmath226 for all @xmath227	in this paper , we consider the cox ingersoll ross ( cir ) process in the regime where the process does not hit zero . we construct additive and multiplicative discrete approximation schemes for the price of asset that is modeled by the cir process and geometric cir process . in order to construct these schemes , we take the euler approximations of the cir process itself but replace the increments of the wiener process with iid bounded vanishing symmetric random variables . we introduce a `` truncated '' cir process and apply it to prove the weak convergence of asset prices . we establish the fact that this `` truncated '' process does not hit zero under the same condition considered for the original nontruncated process . ./style / arxiv - vmsta.cfg cox ingersoll ross process , discrete approximation scheme , functional limit theorems 60f99,60g07,91b25
the discovery of iron - based superconductors @xcite ( fesc ) put forward experimental and theoretical efforts to understand the reason for rather high critical temperatures and symmetry of superconducting order parameters in these compounds . these studies yielded a comprehensive experimental description of the electronic fermi surface structure , which includes multiple fermi surface sheets in a good agreement with density functional calculations @xcite . the fermi surface of the moderate doped fesc is given by two small hole pockets around @xmath2 point and two electron pockets around @xmath3 point in the folded zone . this band structure suggests strong antiferromagnetic fluctuations , which may be a mechanism for electron pairing . in this case , the natural order parameter for most of the fesc is so called @xmath4 state , described by the nodeless order parameter with different signs for electron and hole - like pockets . this model agrees well with the experimentally found for most of the moderately doped fe - based superconductors nodeless character of the order parameter @xcite . however a question , whether the order parameter changes its sign by changeover from electron - like to hole - like pockets , is still under discussion . moreover , relative robustness of the superconductors against nonmagnetic impurities led to a suggestion that a more conventional two - band order parameter a uniform sign change ( @xmath5 ) is realized in these systems @xcite . in our previous paper @xcite it was demonstrated that not only superconductors with @xmath5 order parameter but also @xmath4 may be robust against nonmagnetic impurities . therefore the robustness against nonmagnetic impurities can not be considered as a strong argument against @xmath0 order parameter . moreover , it was shown , that there are two types of @xmath0-superconductors with respect to disorder @xcite . in the first one @xmath6 goes down as disorder is increased , until it vanishes at a critical value of the scattering rate . this behavior is similar to the famous case of abrikosov - gorkov magnetic impurities . it is widely discussed in the literature . in the second type of @xmath0 superconductors @xmath7 tends to a finite value as disorder is increased @xcite ; at the same time the gap functions for the electron - like and hole - like fermi surfaces acquire the same signs , i.e. the transition occurs from @xmath0 to @xmath5 . in the present paper we discuss how the disorder induced transition @xmath8 can display itself in single particle properties and optical conductivity . it is shown that disorder dependence of these characteristics at the transition point is strongly nonmonotonic function of the impurity scattering rate and can be easily seen in tunneling spectroscopy , arpes and optics . therefore systematic study of disorder effects by means formally phase insensitive techniques , mentioned above , may provide information about the sign of the underlying order parameter in the clean limit . the paper is organized as follows . in the section ii we discuss the approximation used for the calculations . the section iii is devoted to single particle properties . we discuss how the transition @xmath8 manifests itself in arpes and tunneling spectroscopy . in the section iii we calculate the two - particle response function and discuss peculiarities which can be seen in optical conductivity in the vicinity of the transition point . the article is concluded with discussions . for the calculations we employ the standard approach of quasiclassical ( @xmath9 - integrated ) green functions in nambu and band space @xcite : @xmath10 with band quasiclassical green functions @xmath11 where the @xmath12 denote pauli matrices in nambu space and @xmath13 is the density of states on the fermi level in the band @xmath14 ( for the sake of simplicity the two band model is considered ) . the function @xmath15 is related to the full green function @xmath16 by the standard procedure of @xmath9-integration @xmath17 . the quasiclassical green functions are obtained by numerical solution of the eliashberg equations : @xmath18 . \ee for simplicity we use the same normalized spectral function of electron - boson interaction $ b(\omega)$ for all the channels , which is presented in inset fig . the maximum of the spectra is $ \omega_{sf } = 18$~mev \cite{charnukha}. the matrix elements $ \lambda^{\tilde\phi}_{\alpha\beta}$ are positive for attractive interactions and negative for repulsive ones . the symmetry of the order parameter in the clean case is determined solely by the off - diagonal matrix elements . the case $ sign \lambda^{\tilde\phi}_{a b } = sign \lambda^{\tilde\phi}_{b a } > 0 $ corresponds to $ { { s_{++}}}$ superconductivity and $ sign\lambda^{\tilde\phi}_{a b } = sign \lambda^{\tilde\phi}_{b a } < 0 $ to $ { { s_{\pm}}}$. the matrix elements $ \lambda^{\tilde\omega}_{\alpha\beta } $ have to be positive and are chosen $ \lambda^{\tilde\omega}_{\alpha\beta } = \|\lambda^{\tilde\phi}_{\alpha\beta}\|$. for further calculations we use the same matrix $ \lambda^{\tilde\phi}_{aa}=3 $ , $ \lambda^{\tilde\phi}_{bb}=0.5 $ , $ \lambda^{\tilde\phi}_{ab}=-0.2 $ , $ \lambda^{\tilde\phi}_{ba}=-0.1 $ for $ { { s_{\pm}}}$-case and $ \lambda^{\tilde\phi}_{aa}=3 $ , $ \lambda^{\tilde\phi}_{bb}=0.5 $ , $ \lambda^{\tilde\phi}_{ab}=0.2 $ , $ \lambda^{\tilde\phi}_{ba}=0.1 $ for $ { { s_{++}}}$-case . the correspondent ratio of the densities of states is $ n_a / n_b = 0.5 $ ( see \cite{spm2spp } ) . the second terms in the right side of the eqs.(\ref{eq.elias.1}-\ref{eq.elias.2 } ) reflect scattering on impurities . in the general case $ \gamma_{\alpha \beta}(\omega)$ can be written in the following form : \be \gamma_{\alpha \beta } ( \omega ) = \gamma^n_{\alpha \beta } i(\omega ) , \label{eq.gamma } \ee where the $ \gamma^n_{\alpha \beta}$ are inter- and intra - band impurity scattering rates in the normal state . the dynamical part $ i(\omega ) = 1 $ in born approximation ( see appendix ) . beyond the born approximation it reads : \be i(\omega ) = \frac{1}{1 - 2 \zeta c_{ab}(\omega ) } , \ee where $ c_{ab}(\omega)$ is the { \it coherence factor } : \be c_{ab}(\omega ) = 1 - \frac{\tilde\omega_{a}\tilde\omega_{b } - \tilde\phi_a\tilde\phi_b } { \sqrt{\tilde\omega_a^2 - \tilde\phi^2_a } \sqrt{\tilde\omega_b^2 - \tilde\phi^2_b } } . \ee note that in the normal state $ c_{ab}(\omega ) = 0 $ and $ i(\omega ) = 1$. the dimensionless constant $ \zeta$ is related to the inter - band impurity scattering rate in the normal state $ \gamma^n_{ab}$ as \be \gamma^n_{ab } = \frac{n_{imp}}{\pi n_a } \zeta . \ee the dependence of $ \gamma^n_{\alpha , \beta}$ and $ \zeta$ on the scattering potential is shown in the appendix . \section{quasiparticle properties } \subsection{dos in superconductive state . } interband scattering is expected to modify the gap functions and the tunneling density of states ( dos ) in the superconducting state in a multiband superconductor . in the weak coupling regime the impurity effects have been discussed in \cite{schopohl } within the born limit and extended in \cite{dolgov2005 } to the strong coupling case . in the following we will calculate the gap functions , and the superconducting dos by solving the nonlinear eliashberg equations in the $ s_{\pm}$ and $ s_{++}$ superconductors for various values of the interband nonmagnetic scattering rate , going beyond the born approximation . total dos in superconducting state is given by the following expression \begin{equation } n(\omega)=\sum_{\alpha}n_{\alpha}(0)\mathop{\rm re}\frac{\omega}{\sqrt { \omega^{2}-\delta_{\alpha}^{2}(\omega+i\delta ) } } , \label{eq : dos}\end{equation } where we have introduced the\textit { complex order parameter } : \be\delta_{\alpha } ( \omega+i\delta)\equiv\omega\tilde{\phi}_{\alpha}(\omega+i\delta ) /\tilde{\omega}_{\alpha}(\omega+i\delta)=$ \textit{re}$\delta_{\alpha}(\omega)+i$ \textit{im}$\delta_{\alpha}(\omega).\ee the solution for $ \delta_{\alpha}(\omega)$ allows the calculation of the current - voltage characteristic $ i(v)$ and \textit{tunnelling conductance } $ g_{ns}(v)=di_{ns}/dv$ in the superconducting state of the $ nis$ tunneling junction . in contrast to a single band case , where dos does not depend on \textit{nonmagnetic } impurities in the multiband case $ \delta_{\alpha}(\omega+i\delta)$ and dos are strongly dependent on interband impurity scattering . figure 1 shows the calculated gap functions $ \delta_\alpha(\omega)$ for the bands $ a$ and $ b$ for different interband impurity scattering rates . one sees in both $ s_{\pm}$ and $ s_{++}$ cases strong non - monotonic frequency dependence of the gap function with the maximums of the absolute values around $ 250 cm^{-1}$ for $ a$-band $ 140 cm^{-1}$ for $ b$-band , originated from the strong electron - boson coupling . furthermore , the effects of impurity scattering are visible as additional structure at low energies comparable to the interband scattering rate $ \gamma_a$. the most spectacular effect is the impurity - induced sign change of $ re \delta(\omega)$ at low energies in $ b$-band in the $ s_{\pm}$ state . this $ s_{\pm } \to s_{++}$ transition was predicted in matsubara representation in our earlier paper \cite{spm2spp } , where its consequences for non - vanishing $ t_c$ vs disorder behavior were discussed . \begin{figure } [ ptbh ] \begin{center } \includegraphics[width=\linewidth ] { fig1.eps } \caption{superconducting gap functions for bands $ a$ and $ b$ at various interband impurity scattering rates in $ s_{\pm}$ and $ s_{++}$ models . the parameters are $ \zeta\approx0.2 $ , $ \gamma^n_{bb}=2\gamma^n_{aa } = \gamma^n_{ab}=2 \gamma^n_{ba } \approx 0.4 \gamma_a $ . they corresponds to scattering strength $ \sigma=0.5$. the relation between $ \sigma$ and $ \gamma_{a}$ to the scattering potential is given in the appendix . } \label{fig1 } \end{center } \end{figure } \begin{figure } [ ptbh ] \begin{center } \includegraphics[width=\linewidth ] { fig2a.eps } \caption{total density of states for various impurity scattering rates in $ s_{\pm}$ and $ s_{++}$ models at low temperature $ t\ll t_{c0}$. the parameters are the same as in fig . 1 . inset shows the electron - boson interaction function $ b(\omega)$. } \label{fig2 } \end{center } \end{figure } figure 2 shows comparison of dos in $ s_{\pm}$ and $ s_{++}$ states for different magnitudes of the interband scattering rate $ \gamma_a$ at low temperatures $ t \ll t_{c0}$. in the clean limit , one sees two different excitations gaps for the two bands . in accordance with earlier calculations for $ s_{++}$ superconductors \cite{dolgov2005 } , the interband impurity scattering mixes the pairs in the two bands , so that the states appear in the $ a$-band at the energy range of the $ b$-band gap . these states are gradually filled in with increasing scattering rate . at the same time the minimal $ b$-band gap in the dos raises due to increased mixing to the $ a$-band with strong electron - boson coupling . in the $ s_{\pm}$ superconductor , the modifications of low - energy dos with interband impurity scattering is completely different . due to sign change of $ re \delta(\omega)$ in the $ b$-band , gapless region exists in a range of values of scattering parameter $ \gamma$ around 35 $ cm^{-1}$ , as clearly seen in the left panel in fig.2 . such gapless behavior manifests itself in optical properties of $ s_{\pm}$ superconductors , as will be demonstrated below . \subsection{arpes and the self - energy } angle - resolved photo emission spectroscopy ( arpes ) probes the photoemission current $ i(\mb{k},\omega)$ , which can be calculated as : \ [ i(\mathbf{k},{{\omega } } ) = \sum_\alpha\|m_\alpha(\mathbf{k},{{\omega } } ) \|^{2}f({{\omega } } ) a_\alpha(\mathbf{k},{{\omega } } ) . \ ] here $ m(\mathbf{k},{{\omega } } ) $ is the dipole matrix element that depends on the initial and final electronic states , incident photon energy and polarization , $ f({{\omega } } ) $ is the fermi distribution function and \be \fl a_\alpha(\mathbf{k},\omega ) = - \frac{1}{2 \pi } \tr\left\ { i m \hat g_\alpha(\mathbf{k},\omega ) \hat\tau_0\right\ } = - \frac{1}{\pi } i m \frac { \tilde{\omega}_\alpha(\omega ) } { \tilde\omega_\alpha^2(\omega)-\xi_{\alpha}^2 ( \mb{k } ) - \tilde\phi_\alpha^2(\omega ) } \label{eq.spectral.function } \ee is single particle response function . in the weak coupling limit the contribution of the electron - boson interaction to self - energy $ \sigma_{\alpha0}(\mb{k},\omega)$ \ ( see first terms in l.h.s . of eqs . \ref{eq.elias.1 } , \ref{eq.elias.2 } ) vanishes . it means , that in the model with isotropic self - energy $ \sigma _ { \alpha 0}^{e - b}(\omega ) \rightarrow 0 $ , $ \sigma _ { \alpha 1}^{e - b}(\omega ) \rightarrow \delta _ { \alpha } ( \omega ) $ . then the single particle spectral function takes the form : \ [ a_{\alpha } ( \mb{k},\omega)=\frac{1}{\pi } im\frac{\omega \left [ 1 + i\sum\limits_{\beta = a , b } \gamma _ { \alpha \beta } /\sqrt{\omega ^{2}-\delta _ { \beta } ^{2}(\omega ) } \right ] } { d } \]with $ $ d=\xi _ { \alpha } ^{2}(\mb{k})+ \left ( \sum\limits_{\beta = a , b}\gamma _ { \alpha \beta } \right ) ^{2}-\omega ^{2}+\delta _ { \alpha } ^{2}(\omega ) -2i\sum\limits_{\beta = a , b } \gamma _ { \alpha \beta } \frac { \left [ \omega ^{2}+\delta _ { \alpha } ( \omega ) \delta _ { \beta } ( \omega ) \right ] } { \sqrt{\omega ^{2}-\delta _ { \beta } ^{2}(\omega ) } } . $ $ in the gaped regime $ a_\alpha(\mb{k } , \omega)$ vanishes below $ \delta _ { \beta } $ but in the gapless one for $ b$-band is the same as in the normal state : \ [ a_{b } ( \mb{k},\omega)=\frac{1}{\pi } im\frac{\omega\left [ 1 + i\sum\limits_{\beta = a , b}\gamma _ { b \beta } /\|\omega\|\right]}{\xi _ { b } ^{2}(\mb{k})-\omega^2 \left(1+i \sum\limits_{\beta = a , b}\gamma _ { b \beta } /\|\omega\| \right ) ^{2}}. \ ] \begin{figure } [ ptbh ] \begin{center } \includegraphics[width=0.9\linewidth ] { fig3a.eps } \caption{the quasi - particle spectral function $ a_{b}(\mathbf{k},\omega)$ \ for $ b$-band with a small gap in the clean limit . the parameters are the same as in fig . 1 . } \label{fig3a } \end{center } \end{figure } the quasi - particle spectral function $ a_{b}(\mathbf{k},\omega ) $ given by eq.(\ref{eq.spectral.function } ) for $ b$-band is shown in fig.3 . in this case the behavior of $ a_{b } ( \mathbf{k},\omega ) $ at small $ \omega $ and $ \xi $ reflects the existence of well - defined energy gap . in contrast to that , the function $ a_{b}(\mathbf{k},\omega ) $ in the regime of $ s_{\pm } \rightarrow s_{++}$ transition shows no gap , as seen from fig.4 . with further increase of scattering rate $ \gamma _ { a}$ , when $ s_{++}$ state is realized , in the $ b$-band energy gap appears again . therefore , arpes measurements at various impurity concentrations may provide useful tool to distinguish underlying pairing symmetry of superconducting state in pnictides . \begin{figure } [ ptbh ] \begin{center } \includegraphics[width=0.9\linewidth ] { fig4a.eps } \caption{the quasi - particle spectral function $ a_{b}(\mathbf{k},\omega)$ for $ b$-band with a small gap in the gapless regime ( $ \gamma_a=40 cm^{-1}$ ) . the parameters are the same as in fig . 1 . } \label{fig4a } \end{center } \end{figure } \section{optical conductivity } the optical conductivity in the london ( local , $ \mathbf{q}\equiv0 $ ) limit in { \it a - b } plane is given by \begin{equation } \sigma(\omega)=\sum_{\alpha}\omega_{pl,\alpha}^{2}\pi_{\alpha}(\omega)/4\pi i\omega , \label{sig}\end{equation } where $ \pi_{\alpha}(\omega)$ is an analytical continuation to the real frequency axis of the polarization operator ( see , e.g. refs . \cite{nam},\cite{lrz},\cite{dgs},\cite{ac},\cite{mars1 } ) \ [ \pi_{\alpha}(\omega)=\left\ { i\pi t\sum_{n}\pi_{\alpha}(\omega'_{n},\nu _ { m})\right\ } _ { i\nu_{m}\longrightarrow\omega+i0^{+ } } , \ ] $ \alpha = a , b$ is the band index . \begin{eqnarray } \fl \pi _ { \alpha}(\omega ) = \int d\omega ^{\prime } \left\ { \frac{\tanh \left ( \frac{\omega _ { -}}{2t}\right ) } { d^{r } } \right . \left [ 1-\frac{\tilde{\omega}_{-}^{r}\tilde{\omega}_{+}^{r } + \tilde\phi _ { -}^{r}\tilde\phi _ { + } ^{r } } { q_{-}^r q_{+}^r } \right ] \label{real } -\frac{\tanh \left ( \frac{\omega _ { + } } { 2t}\right ) } { d^{a}}\left [ 1-\frac{\tilde{\omega}_{-}^{a}\tilde{\omega}_{+}^{a}+ \tilde\phi _ { -}^{a}\tilde\phi _ { + } ^{a } } { q_{-}^a q_{+}^a } \right ] \nonumber \\ -\frac{\tanh \left ( \frac{\omega _ { + } } { 2t}\right ) -\tanh \left ( \frac{\omega _ { -}}{2t}\right ) } { d^{a}}\left . \left [ 1-\frac{\tilde{\omega}_{-}^{a}\tilde{\omega}_{+}^{r}+ \tilde\phi _ { -}^{a}\tilde\phi _ { + } ^{r } } { q_{-}^a q_{+}^r } \right ] \right\ } , \end{aligned}\ ] ] where @xmath19 @xmath20and @xmath21@xmath22 , and the index @xmath23 corresponds to the retarded ( advanced ) brunch of the complex function @xmath24 ( the band index @xmath25 is omitted ) . in the normal state the conductivity is @xmath26 if the dominant contribution to the quasiparticle damping comes from the impurity scattering , it reduces to the drude formula : @xmath27 with @xmath28 in the fig . [ fig5 ] we demonstrate the impact of disorder on the optical conductivity @xmath29 . in the clean limit one sees @xmath30 for @xmath31 . with increase the impurity scattering rate the region @xmath32 for the band @xmath33 decreases and the peak above @xmath34 becomes sharper . it is clearly seen that in the vicinity of transition from the @xmath0 to @xmath5 state ( @xmath35 ) , the conventional drude - response characteristic in the weak for a normal metal - state is realized . the origin of this effect is gapless nature of superconductivity near the impurity - induced @xmath36 transition . with further increase of the impurity scattering rate , the optical conductivity recovers gapped - like behavior , but with smaller gap . it is strikingly different from the behavior of superconductors with @xmath5 order parameter ( fig . [ fig5].b ) , where values of two gapes tends to merge at the limit of infinite impurity scattering rate . this reentrant behavior of optical conductivity with concentration of nonmagnetic impurities may serve as unambiguous indication for the @xmath4 order parameter symmetry . [ ptbh ] [ ptbh ] another important characteristic of the superconducting state is the real part of the electromagnetic kernel ( polarization operator ) which is related to the imaginary part of optical conductivity @xmath37 ( see eq.[sig ] ) . fig.6 shows the frequency dependence of re @xmath38 for @xmath4 and @xmath5 models for various interband scattering rates . one can see that in the @xmath5 case dips at @xmath39 occur for nonzero scattering , in accordance with previous calculations done for single - band superconductors @xcite . further , an interesting peculiarity is seen in the response of the @xmath33-band in the @xmath4 state : the dip position is a nonmonotonic function of interband scattering rate and the dip vanishes completely in the gapless regime corresponding to the @xmath40 transition . magnetic field penetration depth @xmath41 in the local ( london ) limit in _ a - b _ plane is related to @xmath37 in zero frequency limit @xmath42 where @xmath43 is the velocity of light . if we neglect strong - coupling effects ( or , more generally , fermi - liquid effects ) then for a clean uniform superconductor at @xmath44 we have the relation @xmath45 where @xmath46 is the plasma frequency in different bands . partial contributions to the magnetic field penetration depth can be written as 1/_l^2(t)=re _ = a , b _ _ g(t)-0^z_(,t ) . here the points @xmath47 are determined by the condition for the density of states in the band @xmath48 for superconductors with gap nodes as well as for @xmath49 : @xmath50 ( see , @xcite ) . peculiarities of the penetration depth in the crossover regime from @xmath4 to @xmath5 state have been discussed earlier @xcite . we have studied the effects of the impurity - induced @xmath1 transition in the density of states , the single particle response function and optical conductivity in a multiband superconductors with @xmath0 symmetry of the order parameter . it has been shown that smaller gap vanishes in the vicinity of this transition , leading to gapless nature of photoemission and tunneling spectra . in optical response , the @xmath51 transition leads to `` restoring '' of the drude - like frequency dependence of @xmath29 . we have also found interesting anomalies in the real part of polarization operator , with reentrance behavior of the dip - like structure at @xmath52 as a function of interband scattering rate . this effect leads to non - monotonic behavior in the magnetic field penetration depth as a function of the impurity concentration . we want to stress that systematic study of the impact of disorder on the single - particle response function and optical conductivity may give an information about underlying symmetry of the superconductive order parameter . the authors are grateful to p. hirschfeld , m.m . korshunov , a. charnukha , a.v . boris , b. keimer for useful discussions . the present work was partially supported by the dfg priority programme spp1458 ( dve ) , dutch fom ( aag ) . the impurity part of the self energy is obtained by analytic continuation of the correspondent self - energy on matsubara frequencies @xmath53 derived in @xcite . the last is taken in the @xmath54 - matrix approximation : @xmath55 where @xmath56 and @xmath57 is the impurity concentration and the scattering potential : @xmath58 the quasiclassical nambu green s functions on matsubara frequencies are : @xmath59 solutions of eqs . ( [ eq.tmatrix ] ) for @xmath60 and @xmath61 are @xmath62 where @xmath63 [ ptb ] the analytical continuation on the real axis leads to the following expression : @xmath64 and @xmath65 with @xmath66 given by eq.([eq.gamma ] ) . the dimensionless constant @xmath67 is convenient to express in terms of dimensionless scattering potentials @xmath68 and @xmath69 . then it has the following compact form : = . [ eq.app.zeta ] normal state impurity scattering rate reads : ^n _ = [ eq.app.gamma.aa ] and for @xmath70 ^n _ = . [ eq.app.gamma.ab ] the born approximation corresponds to @xmath71 . then up to quadratic terms in @xmath72 one gets : @xmath73 and @xmath74 it is worth to introduce the parameters @xmath75 and @xmath76 . the parameter @xmath77 is used as an indicator of the strength of the impurity scattering . in the born approximation @xmath78 , while in the opposite unitary limit @xmath79 . 99 kamihara y _ et al . _ 2008 _ j. am . soc . _ * 130 * 3296 see , e.g. " superconductivity in iron - pnictides in 2008 _ physica _ c issues 9 - 12 * 469 * ; hirschfeld p j , korshunov m m and mazin i i 2011 _ rep . phys . _ * 74 * 124508 ; kordyuk a a 2012 _ fiz . temp . _ * 38 * , 1119	we investigate effects of disorder on the density of states , the single particle response function and optical conductivity in multiband superconductors with @xmath0 symmetry of the order parameter , where @xmath1 transition may take place . in the vicinity of the transition the superconductive gapless regime is realized . it manifests itself in anomalies in the above mentioned properties . as a result , intrinsically phase - insensitive experimental methods like arpes , tunneling and terahertz spectroscopy may be used for revealing of information about the underlying order parameter symmetry .
symmetry under spatial transformations is a simple yet fundamental concept underlying the description of most systems in contemporary physics as the origin of conserved quantities . in quantum theory and wave mechanics , symmetry - induced conservation laws are extended from continuous to discrete transformations in terms of commutation of the corresponding operators with the hamiltonian , which in turn dictates the spatial structure of stationary states . ubiquitous paradigms are reflection ( @xmath0 ) or finite translation ( @xmath1 ) symmetry imposing definite parity or bloch momentum on hamiltonian eigenstates , which then provide the basis for understanding phenomena such as selection rules of level transitions in atoms @xcite or the formation of band structure in crystals @xcite . in intricate lattice systems of intense current interest such as topological insulators @xcite , weyl semimetals @xcite , and lieb lattices @xcite , state localization and transport properties are also predominantly traced back to broken or unbroken spatial symmetries in combination with the operation of time reversal ( @xmath2 ) . the significance of broken spatiotemporal symmetries and has further been addressed for driven quantum systems @xcite in relation to directional transport @xcite . since symmetries are naturally addressed with respect to commutation with the global system hamiltonian , they are usually regarded dichotomously as broken or unbroken . nevertheless , in a composite system the hamiltonian elements may retain a certain spatial symmetry _ locally _ in a subdomain of configuration space although it is broken _ globally _ ; we refer to this case as a ` local symmetry ' ( ls ) . although there is clearly a remnant of the considered symmetry in the corresponding subsystem , it is not evident how to track this information in the eigenstates of the hamiltonian since it generally does not commute with the ls operation . at the same time , lss are actually inherently paramount to a variety of systems such as aperiodic lattices with long - range order @xcite , partially disordered media @xcite , and single complex molecules @xcite . a typical example are also generic solid state nanostructures @xcite where unavoidable impurities and defects ( usually modeled as @xmath0-symmetric ) separate the host crystal into finite @xmath1-symmetric parts . ls may also be present by design in artificial devices like , e.g. , multilayered photonic setups @xcite , acoustic waveguides @xcite , or magnonic systems @xcite , due to restrictions of finiteness or functionality . in fact , even global spatial symmetry is generally rendered local as soon as a system s immediate environment is included in the description . the spatial structure of stationary states in continuous , one - dimensional ( 1d ) , locally symmetric scattering potentials was recently pinpointed within the framework of symmetry - adapted nonlocal currents ( nlcs ) @xcite : at any real frequency , two distinct nlcs are spatially constant within subdomains of local @xmath0 or @xmath1 symmetry . they provide a mapping of field amplitudes between ls related points , thus generalizing the mapping through parity or bloch factors for global symmetry @xcite . the nlcs were further linked to perfectly transmitting states @xcite and proposed as a natural order parameter for spontaneous symmetry breaking in non - hermitian @xmath3-symmetric systems @xcite , as well as observed experimentally in lossy acoustic setups @xcite . the stationary translational invariance of the nlcs suggests that they are governed by a continuity equation incorporating the associated symmetry transform . in the case of global space reflection @xmath0 , a continuous two - point version of the quantum current density is intricately related to the combination of @xmath0 with the operation of time reversal @xmath2 @xcite . a similar bilocal picture of quantum mechanics has also been used in an alternative description of double slit interference @xcite . conservation of the two - point current in 1d continuous @xmath3-symmetric systems can be derived from schrdinger s @xcite , or via gauge invariance of a nonlocal lagrangian @xcite in analogy to local current conservation . global @xmath3 symmetry has gained increased attention since it admits real discrete spectrum for non - hermitian systems @xcite in parametric regions where density loss and gain are balanced @xcite , and has been intensively investigated in wave scattering @xcite . non - hermitian hamiltonians are of particular interest in discrete models , where the coupling between sites constitutes an additional degree of freedom for design and control @xcite . a reliable platform of applicability is provided by photonic waveguide arrays which are well described by an effective discrete schrdinger equation with tight - binding hamiltonian @xcite . discrete arrays are ideal for implementing global or local spatial symmetries , by adjusting the onsite hamiltonian elements or the inter - site hoppings , to study their impact on field state properties . as an example , in a certain class of discrete structures with ` hidden ' symmetries in the state space @xcite , conserved quantities from global operators commuting with the hamiltonian can be identified . partial @xmath3 symmetry along a given coordinate has also been implemented in the form of dimer arrays featuring synchronous bloch - zener oscillations @xcite or synthetic lattices with solitonic excitations @xcite . in discrete models , even the local current becomes a two - point quantity , naturally defined to flow on the ` link ' between two lattice sites @xcite . this raises the question of how to formulate discrete conservation laws for nonlocal quantities entailing different links . in view of ls , such a framework would provide the relation between nonlocal link currents to general lattice symmetry transformations restricted to subdomains of a composite system . in the present work , we combine the concept of local symmetry transformations with that of nonlocal current continuity for discrete generally non - hermitian schrdinger models . we thereby identify the breaking of lss as sources for the nlcs flowing along transformation related links . the nlcs are given in operator form , thus making the extension to multiple links per site straightforward . with photonic waveguide arrays in mind , we apply the framework to 1d tight - binding hermitian and non - hermitian lattice systems with local @xmath4 or @xmath5 symmetry , respectively , with @xmath6 ( inversion ) or @xmath1 ( finite translation ) . the evolution of the associated nonlocal charge is given in terms of the nlcs at the boundaries of locally symmetric domains , based on a symmetry - induced discrete gauss law . together with an alternatively defined nlc , identities are derived which relate the amplitudes of general states at symmetry - transformed sites and connect the nlcs with the local current . for stationary states , the lss thereby induce amplitude mappings within finite domains . the domainwise translation invariance of the nlcs in stationary states is demonstrated for different types of ls , including their generalized analogue for periodically modulated arrays . it is thus demonstrated how the nlcs reveal the hamiltonian lss encoded in arbitrarily irregular eigenstates . finally , the density profile of perfectly transmitting scattering states is shown to be characterized by nlcs in terms of its ls . the paper is organized as follows . in [ sec : transform ] we introduce the concept of lss in terms of the corresponding local permutation matrices and nonlocal site projection operators . section [ sec : discrete_continuity ] is devoted to the discrete nonlocal continuity and the relation of the nlcs and associated nonlocal charge to lss . in [ sec : stat_invariants_mappings ] we address the properties of nlcs in stationary states , derive the symmetry - induced amplitude mapping relations and the connection to the local current , and generalize the nlc to driven arrays . in [ sec : bound_eigenmodes ] we discuss the cases of complete , overlapping , and gapped ls , and provide examples of invariant nlcs in non - hermitian and driven finite arrays . in section [ sec : scattering ] we consider the nlcs in scattering states , derive their relation to perfect transmission and illustrate their invariance in locally symmetric non - hermitian scatterers . section [ sec : conclusions ] concludes the paper and provides perspectives for generalizations and future work . to define the concept and explicit form of a ls operation in a 1d discrete model , we start by considering a single - particle , time - independent hamiltonian @xmath7 represented on @xmath8 localized site excitations @xmath9 , @xmath10 , with onsite potential elements @xmath11 and hopping elements @xmath12 from site @xmath13 to @xmath14 . the temporal evolution of a quantum state @xmath15 is then governed by the schrdinger equation ( se ) @xmath16 where we set @xmath17 . such a discrete model is widely used to describe coherent electron transport in 1d lattice potentials like the ones induced in quantum dot arrays , but also as a finite difference approximation to the continuous se by choosing appropriate ( typically equal ) hopping values . in particular , [ eq : schr ] effectively also describes light propagation in photonic waveguide arrays within the paraxial approximation @xcite for small lateral wave vector components . 1d versions of such a photonic array are schematically depicted in [ fig : setup ] . time @xmath18 is here represented by the longitudinal spatial coordinate @xmath19 along the waveguides , and energy @xmath20 is accordingly represented by the light propagation constant along @xmath19 . onsite potentials @xmath21 correspond to the refractive indices of waveguides , while the hoppings @xmath22 are proportional to the overlap of the light field localized in waveguides @xmath23 and thereby mapped to their distance . remarkable agreement with experimental results is achieved already in the tight - binding approximation @xcite with @xmath24 for @xmath25 , which we will focus on in the following . given the above framework for a discrete 1d schrdinger model , we consider a bijective mapping @xmath26 given by @xmath27 which performs a local inversion ( @xmath0 ) through a center @xmath28 or translation ( @xmath1 ) by a length @xmath29 of sites @xmath30 to sites @xmath31 , where the domain @xmath32 is generally a subset of the total set @xmath33 of all sites . suppose now that the discrete system decribed by the hamiltonian elements @xmath34 remains invariant under this site mapping , that is , @xmath35 so that @xmath36 is a symmetry transformation of the system . we then say that the system possesses a _ local _ symmetry in the union @xmath37 . a ls is thus contrasted with a _ global _ symmetry of the system for which @xmath38 . locally @xmath0- and @xmath1-symmetric photonic arrays are depicted in [ fig : setup](a ) and ( b ) , respectively . the discrete parameters @xmath28 and @xmath29 here refer to the site indexing and not to the continuous coordinate along the array : the center @xmath28 is integer ( half integer ) for odd ( even ) number of sites @xmath39 in @xmath40 , and the integer period @xmath29 is the shift in site number to the right . intersite distances in the physical system are mapped to corresponding hopping element magnitudes in the model , determined by evanescent field overlap between waveguides in photonics @xcite . thus , a local inversion maps a domain @xmath40 to itself , so that @xmath41 , while a local translation shifts @xmath40 by @xmath29 sites , so that @xmath42 where @xmath43 are the rightmost @xmath29 sites in @xmath44 ( see [ fig : setup ] ) for @xmath45 ( the case @xmath46 will be treated later in [ sec : symmetry_types ] ) . ) symmetric about a center @xmath28 or * ( b ) * translation ( @xmath1 ) symmetric with period @xmath29 , for sites @xmath47 within a domain @xmath40 ( solid blue line ) , together with the corresponding close - coupling hamiltonian matrix @xmath48 and local transformation ( site permutation ) matrix @xmath49 or @xmath50 , respectively . @xmath49 maps @xmath40 to itself , while @xmath50 performs a cyclic shift permutation by @xmath29 sites on @xmath51 , where @xmath52 ( solid purple line ) is the rightmost period in @xmath44 ( dotted blue line ) . ] with the ( symmetry ) mapping @xmath36 transforming the site indices @xmath47 , we correspondingly define an operator @xmath53 which replaces the wave amplitude @xmath54 with @xmath55 for the basis ket @xmath9 when acting on a state @xmath56 ( cf . [ eq : psi ] ) , @xmath57 since the basis @xmath58 is assumed orthonormal ( @xmath59 ) , so that we can symbolically write the action as an amplitude mapping @xmath60 in analogy with the local _ site _ mapping in [ eq : fmapping ] for @xmath61 . the associated matrix @xmath62 has elements @xmath63_{mn } = \braket{m\|\hat{\varsigma}\|n}$ ] given by @xmath64_{mn } = \begin{cases } \delta_{\bar{m}n } , ~~&m \in { \mathbb{d}}\\ \delta_{mn } , ~~&m \in { \mathbb{n}}\setminus { \mathbb{u}}\end{cases } , ~~ { { \varsigma}}_{\mathbb{d}}= { { p}}_{\mathbb{d}},{{k}}_{\mathbb{d}}.\ ] ] for local inversion , @xmath49 performs a ( partial ) mirror permutation of the sites within @xmath40 . for local translation , [ eq : sigma_matrix ] leaves the mapping of sites @xmath65 undefined , for which we explicitly choose @xmath63_{mn } = \delta_{\bar{m}-d , n}$ ] so that the last @xmath29 sites in @xmath44 are mapped to the first @xmath29 ones . the permutation matrix thus acts on the whole array but permutes only sites amplitudes within the domain @xmath44 where the ls of the system resides while leaving sites outside @xmath44 unchanged . note that multiple symmetry domains @xmath66 ( @xmath67 ) may in general be present , with corresponding centers @xmath68 or periods @xmath69 . the total symmetry domain is then the ( generally disconnected ) set @xmath70 , with the mapping for each separate @xmath66 entering [ eq : sigma_matrix ] . different types of ls will be addressed in [ sec : symmetry_types ] . symmetry of a system under a transformation is usually associated with the commutation of the corresponding symmetry operator with the hamiltonian . in the case of global symmetry , that is , @xmath38 in the present setting ( with @xmath71 implied for translation symmetry ) , the hamiltonian indeed commutes with the matrix @xmath62 and the energy eigenstates are also eigenstates of the symmetry operation with the corresponding eigenvalues : under inversion or translation the wave amplitudes are mapped to @xmath72 with the parity @xmath73 or the bloch phase @xmath74 ( @xmath75 being the quasimomentum ) as prefactors , respectively . in the case of a ls , although the system remains invariant under the site permutation , the associated operator generally does not commute with the hamiltonian , @xmath76 \neq 0.\ ] ] the reason is the change in connectivity of the domain @xmath40 to its surroundings : the end sites of @xmath40 in a 1d array are not coupled to the same sites after the local permutation , and so the transformed hamiltonian @xmath77 ( recall that @xmath78 for any permutation matrix ) generally contains altered nonzero hopping elements @xmath79 for @xmath80 and @xmath81 ; cf . [ eq : hmn_local_ssym ] . the spatial structure of system eigenstates is therefore no longer dictated directly by the presence of symmetry as in the global case where the field magnitude @xmath82 has the @xmath83-symmetry of the potential . the inequivalence between _ system _ symmetry and commutation of the symmetry operation with the hamiltonian does not come as a surprise : a trivial example is a finite homogeneous tigh - binding array ( with @xmath84 ) which is obviously invariant under _ any _ arbitrary site permutation whose matrix , however , does not commute with @xmath48 ( except for mirror permutation ) , since it changes the intersite connections , as explained above . the question raised here is , nevertheless , if the presence of symmetries in the system are in some way imprinted on the structure of its states , even if those symmetries are not global but of finite extent . the aim is thus to find a characterization of the field amplitude configuration in terms of a quantity which follows the underlying lss of the system . to do so , in the following section we turn to the notion of nonlocal current - density continuity in its discrete form and adapt it to the ls framework . before closing this section , we need to extend the defined local spatial transformations to their combination with the operation of time - reversal @xmath85 , represented here as usual by an antiunitary operator @xmath86 performing complex conjugation @xcite , @xmath87 in the spatial representation . time - reversal thus acts on a state in the basis @xmath58 as @xmath88 and transforms the hamiltonian as @xmath89 ( defined by @xmath90 ) . the latter becomes relevant for complex onsite elements @xmath91 where positive ( negative ) @xmath92 models the gain ( loss ) rate of the state density at site @xmath47 ( as clarified in the next section ) , while complex hoppings @xmath93 with peierls phase @xmath94 generally arise from the presence of gauge fields ( effectively also in 1d systems @xcite ) . if the hamiltonian obeys ( cf . [ eq : hmn_local_ssym ] ) @xmath95 then we will call the system locally @xmath5-symmetric , that is , @xmath3-symmetric or @xmath96-symmetric within the finite spatial domain @xmath97 . note here that we treat ( local ) non - hermiticity , expressed through the @xmath92 , solely as an _ effective _ simple model to express gain or loss of density from or to the surrounding medium , rather than a fundamental description of the physical system . having defined the concept of ls for lattice systems , we shall now generalize the discrete continuity relation @xcite to a nonlocal form adapted to a given the symmetry mapping . in equivalence to the schrdinger equation ( [ eq : schr ] ) for a general state @xmath56 , the temporal evolution of the local density @xmath98 fulfills the discrete continuity equation @xmath99 with @xmath100 being the local density operator in the schrdinger picture , where @xmath101 is the sum of the magnitudes @xmath102 of the currents @xmath103 flowing inwards from sites @xmath104 to site @xmath47 . under the assumption of hermitian hoppings , @xmath105 , the total current magnitude can be written as the expectation value @xmath106 of a current operator @xmath107\ ] ] assigned to site @xmath47 . note that the discrete form of the current divergence entering the continuity equation is generally given by the sum of the projections @xmath108 of the link currents @xmath109 from point @xmath110 to point @xmath111 onto their directions @xmath112 , scaled by the distance @xmath113 @xcite . the local ( referring to site @xmath47 ) continuity equation ( [ eq : continuity_local ] ) shows that the imaginary part @xmath114 of the onsite element @xmath21 acts like a source ( @xmath115 ) or sink ( @xmath116 ) of density @xmath117 at site @xmath47 , in addition to the inflow of current @xmath118 . the partial link current operators producing the @xmath103 can be written as @xcite @xmath119 in terms of the parital velocity operators @xmath120 $ ] with @xmath121 being the position operator , such that the operators producing the projections @xmath122 are given by @xmath123 here , the @xmath124 are the ` hopping operators ' with matrix elements @xmath125 , that is , the upper / lower diagonal of the tridiagonal matrix @xmath48 in the tight - binding case . with respect to the spatial transformation @xmath36 , we now generalize the local discrete continuity equation in [ eq : continuity_local ] to a nonlocal form by replacing the @xmath126 with @xmath127 from [ eq : sigma_operator ] . the latter is now identified with a nonlocal density operator corresponding to the off - diagonal part of the site - represented density operator @xmath128 associated with the transformation @xmath4 , so that @xmath129 consitutes a symmetry - adapted nonlocal density in state @xmath130 , while its sum over @xmath47 yields the total ` nonlocal charge ' @xmath131 also known as ` quasipower ' @xcite in globally @xmath3-symmetric photonic systems , in analogy to the usual charge ( i.e. quantum probability ) , or power in photonics , @xmath132 ( @xmath133 being the identity operator ) . using the se , the temporal evolution of @xmath134 is given by @xmath135 with @xmath136 characterizing the onsite asymmetry , and where @xmath137 assigned to site @xmath47 is the sum of the _ nonlocal _ currents defined by @xmath138 for a local transformation @xmath139 given in [ eq : fmapping ] , with @xmath140 and @xmath141 for @xmath6 ( inversion ) and @xmath1 ( translation ) , respectively . considering hermitian hoppings , @xmath142^\dagger$ ] , these four - site currents @xmath143 and @xmath144 can be written as the expectation values @xmath145 of the corresponding nonlocal current operators @xmath146 in analogy with [ eq : j_n_+-_operator ] , with the transformation type @xmath0 or @xmath1 entering @xmath147 here explicitly indicated . outside the domain @xmath44 of the mapping where @xmath148 , the translation current reproduces the usual current , @xmath149 , and so [ eq : continuity_nonlocal ] reproduces the local current - density continuity in [ eq : continuity_local ] . for both @xmath0 and @xmath1 , the total current operator is given by @xmath150,\ ] ] where @xmath151 is the diagonal part of the hamiltonian . note here that , like the usual currents @xmath122 , the @xmath152 can be seen as link currents , though not flowing along one link but along the two ones corresponding to the hoppings @xmath153 and @xmath154 entering [ eq : q_n+- ] . this nonlocal ` flow ' is illustrated in [ fig : nonlocal_flow ] . using the transformation matrices @xmath155 defined in [ eq : sigma_matrix ] ( and depicted in the example of [ fig : setup ] ) , the @xmath156 in [ eq : q_n+- ] for the whole setup @xmath33 can be written as the components of an @xmath8-entry column vector @xmath157^\top$ ] given by ( cf . [ eq : q_n_+-_p_operator , eq : q_n_+-_k_operator ] ) [ eq : q_n_+-_vector ] @xmath158 p_{\mathbb{d}}h^\pm { { \boldsymbol{\psi } } } - \text{d}[{{\boldsymbol{\psi}}}^\dagger h^\pm ] p_{\mathbb{d}}{{\boldsymbol{\psi } } } , \label{eq : q_n_+-_vector_p } \\ i{{\boldsymbol{q}}}^{\pm}_{\mathcal{{k}}}= \text{d}[{{\boldsymbol{\psi}}}^\dagger ] k_{\mathbb{d}}h^\pm { { \boldsymbol{\psi } } } - \text{d}[{{\boldsymbol{\psi}}}^\dagger h^\mp ] k_{\mathbb{d}}{{\boldsymbol{\psi } } } \label{eq : q_n_+-_vector_k}\end{gathered}\ ] ] for local @xmath159-transformation , respectively , where @xmath160^\top$ ] and @xmath161 $ ] denotes the @xmath162 diagonal matrix with the components of ( row or column ) vector @xmath163 along its diagonal . recall here the convention of mapping the last period @xmath43 of a domain @xmath44 to the first one for @xmath1-transforms . the above expression for the @xmath152 becomes convenient when evaluating them for a given hamiltonian and ( numerically computed ) state amplitudes @xmath54 , as will be done in the following sections . -related links for * ( a ) * inversion @xmath6 and * ( b ) * translation @xmath164 transform . solid ( dotted ) lines indicate the term @xmath165 ( @xmath166 ) entering each upper ( @xmath167 ) and lower ( @xmath168 ) nlc @xmath169 ( @xmath170 ) on the links into a site pair @xmath171 ; see [ eq : q_n+- ] . inset : the translation nlc reproduces the local link current @xmath122 for @xmath148 . ] for a given site transformation @xmath4 , the form of the nonlocal continuity equation ( [ eq : continuity_nonlocal ] ) is similar to [ eq : continuity_local ] with the loss / gain function @xmath92 replaced by @xmath172 which characterizes the sources for the nonlocal density @xmath134 . for a system with @xmath4-symmetric onsite potential , @xmath173 or @xmath174=0 $ ] , we have @xmath175 and the sources ( sinks ) of @xmath134 are represented by the gain ( loss ) of @xmath117 at individual sites @xmath47 like in [ eq : continuity_local ] , and @xmath176/i$ ] like in [ eq : j_n_tot_operator ] . those sources and sinks vanish if the potential is real , but if the @xmath21 are not @xmath4-symmetric , then in [ eq : continuity_nonlocal ] there is a contribution to @xmath172 from the potential differences between @xmath4-related sites . in other words , locally _ broken spatial potential symmetries act as sources _ for the associated nonlocal density . if a system has locally @xmath5-symmetric non - hermitian onsite potential , @xmath177 or @xmath178=0 $ ] , then @xmath179 within the symmetry domain @xmath40 and [ eq : continuity_nonlocal ] reduces to @xmath180 . this means that the net nlc @xmath181 vanishes within @xmath40 for stationary states , in contrast to @xmath118 which may vary spatially for @xmath182 ; see [ eq : continuity_local ] . an alternative nonlocal continuity equation can finally be established by projecting @xmath183 onto the time - reversed state @xmath184 instead of @xmath56 in [ eq : nonlocal_density ] , yielding a _ nonlocal ` density ' @xmath185 using the se , the evolution of @xmath186 can be shown to obey the continuity equation @xmath187 now with asymmetry function @xmath188 and with bitemporal net nlc @xmath189 flowing into site @xmath47 , where @xmath190 defined in similarity to the nlcs in [ eq : q_n+- ] . under the assumption of _ equidirectional _ hoppings , so that @xmath191^\top$ ] , the @xmath192 are given by @xmath193 which resembles an expectation value as in [ eq : q_n_+-_expval ] , though with projection onto @xmath194 . the potential asymmetries @xmath195 again act like sources for the alternative nlc , with @xmath196 for real @xmath21 . in contrast to [ eq : continuity_nonlocal ] , however , an @xmath5-symmetric potential now yields @xmath197 , with the sources ( sinks ) for @xmath117 representing sinks ( sources ) for @xmath186 . an @xmath4-symmetric potential ( @xmath198 ) , on the other hand , leads to vanishing @xmath199 for energy eigenstates within a symmetry domain @xmath40 , even in the presence of unbalanced loss and/or gain where @xmath200 may vary spatially . the relation of _ stationary _ nlcs to lss will be elaborated on in the next section . to see first the impact of the system symmetry on a general ( nonstationary ) state @xmath56 via the nlcs , let us consider the evolution of the nonlocal charge ( see [ eq : nonlocal_charge ] ) within a subdomain @xmath32 , denoted @xmath201 , which obeys ( cf . [ eq : continuity_nonlocal ] ) @xmath202 where @xmath203 is the total net nlc within @xmath40 . now , if the _ hoppings _ of the system are @xmath204-symmetric , @xmath205 , then from the nlc expression ( [ eq : q_n+- ] ) we get , in general , @xmath206 where we define the _ hopping ratio _ @xmath207 between the hopping elements in opposite directions on the link @xmath208 , here with @xmath209 . for equidirectional hoppings within @xmath40 we have @xmath210 ( as is the case for real hermitian hoppings ) , so that @xmath211 equals the sum @xmath212 of currents @xmath213 on links above / below sites @xmath214 of the upper / lower boundaries @xmath215 of @xmath40 , @xmath216 in the form of a nonlocal discrete gauss law . hence , if also the potential is @xmath204-symmetric ( @xmath217 ) , the temporal change @xmath218 in [ eq : charge_evolution ] is then affected only by those boundary currents @xmath212 . recall that @xmath40 may be disconnected , being a union of distinct subdomains @xmath66 as mentioned above ( after [ eq : sigma_matrix ] ) , meaning that it can entail equally many @xmath213 pairs at its boundary points . in the special case of _ global _ @xmath3-symmetry in @xmath38 , we have @xmath179 but also @xmath219 since the boundary currents @xmath220 vanish for the isolated array ( where @xmath221 by definition ) . thus @xmath222 in [ eq : q_d ] and thereby @xmath223 , reproducing the well - known constancy of the nonlocal charge ( quasipower ) in systems with @xmath3-symmetric potential @xcite . with the substitution @xmath224 , the above considerations can be equally applied to the evolution of a bitemporal nonlocal charge @xmath225 . this is then constant in time for a globally @xmath0-symmetric ( non - hermitian ) array @xmath38 with unit hopping ratios where , e.g. , only lossy elements @xmath116 are present . having established the concept of nonlocal discrete continuity in the above operator form and adapted it to symmetry mappings , we will now use the nlcs to show how lss of the lattice system are encoded into its stationary states . the site amplitudes @xmath226 of an eigenstate @xmath227 of the hamiltonian with eigenvalue @xmath20 , @xmath228 are given by the system of close - coupling equations @xmath229 where the set of sites @xmath33 covers the system of interest : for an isolated system closed boundary conditions @xmath230 are imposed by setting @xmath231 , while a scattering setup is realized by appending homogeneous semiinfinite chains with @xmath232 and @xmath233 on the left ( @xmath234 ) and right ( @xmath235 ) of the given array ( see [ sec : scattering ] ) . if the eigenenergy @xmath20 is real , then the state @xmath236 is _ stationary _ in the sense that its local density @xmath237 does not change in time . for a complex eigenenergy @xmath238 ( @xmath239 ) of a non - hermitian @xmath7 , however , the density increases ( decreases ) exponentially in time as @xmath240 for positive ( negative ) @xmath241 and so the state can be considered as _ nonstationary_though not containing components of distinct frequencies . in this general case of complex @xmath20 , the nonlocal continuity equation ( [ eq : continuity_nonlocal ] ) for @xmath236 yields @xmath242 thus , at sites with @xmath5-symmetric onsite elements ( @xmath179 ) , a stationary state @xmath236 with @xmath243 has vanishing sum of nlc amplitudes , @xmath244 following a nonlocal form of a quantum kirchhoff law . for @xmath5-symmetric hoppings @xmath245 , the above equation can be combined with [ eq : q_n-1_+_equal_-q_n_- ] to obtain @xmath246 for consecutive nonlocal link currents . thus , in a _ connected _ domain @xmath40 where the system has @xmath5-symmetric hamiltonian elements , obeying [ eq : hmn_local_stsym ] , the nlc @xmath247 ( @xmath248 ) accumulates the hopping ratios as a prefactor when traversing the domain forwards ( backwards ) . if the hoppings are complex and _ hermitian _ , then @xmath249 and the currents accumulate a peierls - like hopping phase factor along the path on the array . if , additionally , the hoppings are real as is the usual case in 1d setups ( or generally for complex @xmath250 , which is unusual for quantum models but relevant for , e.g. , photonic arrays @xcite , depending on the host medium ) , the @xmath169 are translationally _ invariant_that is , spatially constant in @xmath40 for the stationary state . for the case of @xmath4-symmetry , an analogous local invariance can be derived for the two bitemporal nlcs @xmath192 . the associated continuity equation ( [ eq : continuity_nonlocal_bitemporal ] ) reduces to @xmath251 for an energy eigenstate @xmath236 , so that the kirchoff law @xmath252 holds in a domain @xmath40 with @xmath4-symmetric potential , that is , with @xmath253 . if also the hoppings are @xmath4-symmetric in @xmath40 , then @xmath254 from [ eq : qtrev_n+- ] , yielding @xmath255 for @xmath256 , in analogy to [ eq : q_n-1_+_equal_-q_n_- ] . thus , for equidirectional hoppings ( @xmath257 ) the @xmath192 are translation invariant within the @xmath4-symmetric domain @xmath40 , while accumulating a peierls - like phase along it for complex hermitian hoppings . note that the above applies also to complex eigenenergies @xmath20 , since the ` width ' @xmath241 is absent in [ eq : continuity_nonlocal_bitemporal_eigenstate ] , and thereby to general _ non - hermitian _ arrays with locally @xmath4-symmetric onsite loss / gain elements . to establish a connection between the spatial transformations and the structure of the wave amplitudes via nlcs ( see [ sec : current_mapping ] below ) , it is finally convenient to introduce an equal - time alternative to the bitemporal nlc @xmath192 , defined in similarity to [ eq : q_n+- ] by @xmath258 simply with @xmath259 replacing their complex conjugates in @xmath169 , where @xmath260 . for equidirectional hoppings , they can be expressed via the nlc operators as @xmath261 by projecting onto the complex conjugated state @xmath262 ( that is , without the time reversal @xmath263 applied for @xmath194 ) . in contrast to the @xmath169 and @xmath192 , the @xmath264 are time - dependent even for a stationary state ( with @xmath243 ) through the harmonic factor @xmath265 , and do not obey a nonlocal continuity equation like [ eq : continuity_nonlocal , eq : continuity_nonlocal_bitemporal ] . their conditional _ spatial _ constancy for an energy eigenstate @xmath236 in an @xmath4-symmetric domain @xmath40 can , nevertheless , be derived directly noticing that @xmath266 by using the time - independent schrdinger eigenvalue equation ( [ eq : schr_cc ] ) expressed at transform - related points @xmath47 and @xmath267 . the ls condition ( [ eq : hmn_local_ssym ] ) together with [ eq : schr_diff ] and [ eq : qtilde_n+- ] then lead to @xmath268 , just like in [ eq : qtrev_n_+-_transfer ] . indeed , since @xmath269 for an energy eigenstate , it will have the same spatial characteristics as the stationary @xmath270 . we therefore focus on the equal - time nlcs @xmath271 and @xmath272 in the remainder of the work . in the following , we list some properties of ( stationary ) nlcs and demonstrate their connection to the local current @xmath122 and their use in amplitude mappings between symmetry - related sites in finite domains . the ` transfer ' relations in [ eq : q_n_+-_transfer , eq : qtrev_n_+-_transfer ] for the nlcs of stationary states apply quite generally for @xmath274 and @xmath275 symmetry , respectively , for a connected domain @xmath32 . we now consider their translation invariance in the concrete case of equidirectional hoppings @xmath276 , with _ real _ hermitian hoppings being naturally relevant for most 1d arrays , and distinguish some basic properties as follows : + * ( i ) * in a stationary state @xmath236 with real energy @xmath20 , local @xmath277-symmetry ( @xmath278 , @xmath279 ) or local @xmath280-symmetry ( @xmath281 , @xmath282 ) yields spatially constant corresponding currents @xmath283 for all @xmath30 with @xmath276 , the explicit form of the @xmath284 here being given by @xmath285 respectively . in particular for @xmath277 symmetry it can be shown that @xmath286 vanishes for a state vector @xmath287 whose part @xmath288 is a local @xmath3 ` eigenstate ' , _ and vice versa _ , that is , @xmath289 with unimodular ` eigenvalue ' , @xmath290 , which in turn corresponds to locally symmetric density @xmath117 within @xmath40 . in fact , this applies for arbitrary @xmath277-symmetric hoppings , since no hopping ratio is accumulated in [ eq : q_n_+-_transfer ] for zero current . for the ( local ) @xmath3 eigenstate above it further holds that @xmath291 connecting the alternative nonlocal current with the local one . + * ( ii ) * for local @xmath292-symmetry ( @xmath293 , @xmath294 ) or local @xmath295-symmetry ( @xmath296 , @xmath297 ) , the alternative ( time - dependent ) quantities @xmath264 are spatially constant in @xmath40 , @xmath298 with their explicit form for @xmath236 given by @xmath299 respectively . we underline that those invariants exist also for non - hermitian @xmath36-symmetric potentials , e.g. with ( locally ) symmetric loss profile or ` unbalanced ' loss / gain elements . in analogy to [ eq : q_p_zero ] , @xmath300 vanishes iff the part @xmath301 is a local @xmath0 ` eigenstate ' , @xmath302 now with eigenvalue @xmath303 , and then also @xmath304 holds for the nlc ( recall that generally @xmath305 even for a finite array in the presence of sources and/or sinks ) . again @xmath117 is locally symmetric in @xmath40 for arbitrary @xmath292-symmetric hoppings . note that , for real hermitian @xmath7 , [ eq : q_p_zero , eq : qtilde_p_zero ] are equally applicable for bound arrays and for scattering . an application of the latter in relation to perfect transmission is addressed in [ sec : scattering ] . + * ( iii ) * for a general state @xmath56 , the nlc and its alternative are by construction @xmath306-symmetric and @xmath307-antisymmetric , respectively , @xmath308 noting that the link @xmath309 is @xmath0-mapped to @xmath310 . if the hoppings @xmath245 are equidirectional ( real hermitian in practice ) and @xmath306- or @xmath307-symmetric , then we further have that @xmath311 within @xmath40 , respectively , for _ arbitrary _ onsite elements @xmath21 . thus , the @xmath312 or @xmath313 are then locally symmetric about the center @xmath28 ( that is , about the shifted center @xmath314 if assigned to sites ) for any complex potential profile , regardless of its symmetry . if the potential is @xmath306-symmetric ( @xmath307-symmetric ) , then the first equation above together with the spatial constancy , [ eq : q_constant_d ] , of the @xmath143 for an energy eigenstate @xmath236 with real ( complex ) @xmath20 implies that they are imaginary , @xmath315 . + * ( iv ) * for an eigenstate ( bloch state ) @xmath316 of the finite translation operator @xmath317 with wavefunction @xmath318 , where @xmath319 , the associated nlc equals the local current times the bloch factor while its alternative vanishes , @xmath320 this applies for energy eigenstates of globally @xmath1-symmetric systems of period @xmath29 with symmetry domain @xmath38 , including finite ring structures ( where site @xmath321 is coupled to site @xmath322 ) . recall here that a @xmath96-symmetric potential , where the @xmath144 above may be spatially constant , has period @xmath29 only if it is real ( whereby it is @xmath1-symmetric ) , in which case also the @xmath323 are conserved . + * ( v ) * the nlcs for a forward translation @xmath324 in a general state @xmath56 are related to those for a backward translation @xmath325 from the mapped site @xmath326 by @xmath327 for @xmath96- and @xmath1-symmetric equidirectional hoppings , respectively . thus , for a stationary state ( with real @xmath20 ) , although the @xmath328 are constant only in the subdomain @xmath40 of a locally @xmath96-symmetric domain @xmath42 ( see [ fig : setup ] ) , we can assign @xmath329 to ( links from ) sites @xmath47 in the last period @xmath43 , and similarly for @xmath1 symmetry and the @xmath330 . this way the symmetry is expressed by nonlocal invariants throughout the whole domain @xmath44 . the wave amplitude @xmath331 at the transformed site is related to that of the original one @xmath54 and its conjugate via the @xmath332 through the general identity @xmath333 holding for arbitrary @xmath7 , where both the upper and lower sign in @xmath334 can be chosen ; see derivation in [ app : mapping_derivation ] . for locally @xmath4-symmetric and real equidirectional hoppings @xmath335 ( @xmath30 ) , also the _ neighboring _ amplitudes at @xmath104 and @xmath336 are related via the _ same _ currents ( now with corresponding sign in @xmath334 ) ; see [ eq : current_mapping_matrix_general ] . in other words we have @xmath337 [ \psi_n,\psi_n^ * ] ^\top = [ q_{n \mp 1,{\mathcal{{s}}}}^\pm,-\mathring{q}_{n \mp 1,{\mathcal{{s}}}}^\pm ] [ \psi_n,\psi_n^ * ] ^\top , $ ] so that the @xmath4-transformed amplitudes at @xmath338 are related by _ both _ @xmath334 nlcs ` flowing ' on _ either _ of the attached link pairs @xmath309 , @xmath339 ; see nonlocal ` flow ' in [ fig : nonlocal_flow ] . if , further , the potential @xmath21 is both @xmath204- and @xmath36-symmetric ( that is , real and @xmath36-symmetric ) , then both @xmath340 and @xmath341 , as well as @xmath342 , are spatially constant within @xmath40 for a stationary state @xmath236 . hence , [ eq : current_mapping ] becomes @xmath343 for all @xmath47 in @xmath40 , providing a linear _ mapping relation _ from @xmath344 to @xmath345 with constant coefficients @xmath346 and @xmath347 which can be evaluated for any link @xmath309 in @xmath40 . note that the same stationary mapping of [ eq : current_mapping_stat ] would be arrived at by formulating the identity in [ eq : current_mapping ] using the bitemporal nlc @xmath192 instead of @xmath264 ; see [ app : mapping_derivation ] . -symmetric subdomain ( blue line ) of a 1d array , the amplitude at site @xmath47 is mapped to its image at @xmath326 , as indicated by arrows together with the quantities required for the mapping . * ( a ) * current mapping via [ eq : current_mapping_stat ] in terms of translation invariant local currents @xmath122 and nlcs @xmath152 , @xmath348 evaluated at @xmath349 using @xmath350 and @xmath351 ( red line ) . dotted contour on end sites indicate current flow to ensure @xmath305 . * ( b ) * summation mapping via [ eq : sum_map_pt ] in terms of amplitudes and couplings up to the domain center @xmath28 ( green line ) and the domain invariant @xmath352 . ] this mapping is clearly not applicable for real spatial components @xmath226 of @xmath236 in @xmath40 ( or trivially complex through a common prefactor ) , since the @xmath353 then vanish . at the same time , real @xmath226 yield equimodular @xmath169 and @xmath264 , which in turn coincides quite generally with vanishing local current . indeed , a general identity connecting the nonlocal and local currents on links to sites @xmath47 and @xmath267 is ( see [ app : mapping_derivation ] ) @xmath354 with the sign @xmath355 for @xmath170 , respectively , implying @xmath356 and @xmath357 . this _ current connection _ holds also for @xmath358 , in which case @xmath359 , where @xmath360 is the phase of the general wavefunction @xmath361 , as can easily be checked . for a stationary state we further have that @xmath362 within a hermitian domain @xmath363 . thus , the wave amplitudes in _ different _ local @xmath5 or @xmath4 symmetry domains @xmath66 ( @xmath67 ) in a globally hermitian system are interrelated by common differences @xmath364 through [ eq : current_connection ] for @xmath170 . considering @xmath4-symmetry , the mapping relation ( [ eq : current_mapping_stat ] ) can be seen as a generalization of the parity and bloch theorems ( which map amplitudes to a symmetry - related site via the associated eigenvalue ) from global to local @xmath0- and @xmath1-symmetry , respectively , where breaking of the global symmetry adds an admixture of the complex conjugate @xmath365 to the mapping . taking the continuum limit @xmath366 ( setting all hoppings equal ) , [ eq : current_mapping_stat ] reproduces the generalized bloch and parity theorems of ref.@xcite for stationary states . those are here extended to non - hermitian hamiltonians with variable hoppings and are shown to follow from the general identity ( [ eq : current_mapping ] ) for arbitrary states . the usual ( global ) versions @xmath367 or @xmath368 are recovered directly by inserting eqs.([eq : qtilde_p_zero]),([eq : q_p_j_peig ] ) or [ eq : q_k_bloch ] with @xmath38 for a parity or bloch eigenstate , respectively , into [ eq : current_mapping_stat ] ( or eqs.([eq : q_p_zero]),([eq : qtilde_p_j_pteig ] ) for a @xmath3 eigenstate ) . note that , to avoid @xmath358 for @xmath0 eigenstates , a non - hermitian @xmath0-symmetric system can be considered , subsequently taking the limit of vanishing loss / gain to reproduce the hermitian case . although the identity of the current connection in [ eq : current_connection ] holds for any hamiltonian @xmath7 and any state @xmath56 , the mapping relation of [ eq : current_mapping]and its form for a stationary state within a locally symmetric domain , [ eq : current_mapping_stat]are applicable only for nonzero local current . this is realized in 1d systems for a scattering setting with open boundaries ( connected to asymptotic leads , see [ sec : scattering ] ) through which current flows , or for a finite setup with current sources and/or sinks ( see [ sec : non - hermitian ] ) , where the field amplitudes @xmath369 in stationary states are ( nontrivially ) complex . we now show that the nlcs ( which generally do not vanish even if the local current does ) can be used to map also _ real _ amplitudes of bound hermitian systems between remote sites , thereby exploiting the information of ls . indeed , assuming @xmath3-symmetric hoppings @xmath205 for @xmath30 , we notice that the nlcs @xmath152 for a state @xmath56 can be written ( times @xmath370 ) as @xmath371 with the forward / backward discrete difference operator @xmath372 defined here by @xmath373 for any mapping @xmath374 ( set to @xmath375 or @xmath376 above ) . summing the discrete quotient difference @xcite in [ eq : q_n_disc_quotient ] from site @xmath47 to @xmath377 for a given site @xmath378 and solving for the mapped amplitude @xmath55 we obtain @xmath379 applicable for nonzero @xmath380 . hence , in the case of real amplitudes @xmath226 of a stationary state @xmath236 within a locally @xmath3-symmetric domain @xmath40 with odd number of sites @xmath39 , choosing the fixed point to be the symmetry center @xmath381 ( so that @xmath382 ) simplifies the above expression to @xmath383 where @xmath384 is the symmetry - induced invariant in @xmath40 ( with the upper sign selected here ) . in this manner the image amplitudes @xmath385 can be determined from multiple @xmath226 s ( including the @xmath351 needed to evaluate @xmath386 ) ; see [ fig : mapping ] . the same can be done for an odd - sized @xmath0-symmetric domain , correspondingly using @xmath387 and @xmath388 in [ eq : sum_map_pt ] . in other cases ( e.g. @xmath389 symmetry with even @xmath39 , or @xmath390 symmetry ) , [ eq : sum_map_gen ] yields the amplitude mapping provided that the quotient @xmath391 can be previously determined . the summation mapping presented here holds also for zero current ( implying real @xmath392 for real @xmath226 ) and can thus be seen as a complement to the current mapping in [ eq : current_mapping ] . for a time - dependent hamiltonian @xmath393 ( equivalently , varying along the longitudinal direction @xmath19 of a photonic waveguide array ) , there are no stationary eigenstates and so no spatially invariant nlcs in symmetry domains . in the case , however , of a _ periodic _ modulation with frequency @xmath394 , @xmath395 the schrdinger equation ( [ eq : schr ] ) admits solutions of the form @xmath396 where the real quasienergies @xmath397 ( named in analogy to the quasimomentum of bloch states in spatially periodic systems ) are defined up to multiples of @xmath398 . being _ quasi_-stationary eigenstates of the modified hamiltonian operator @xmath399 with eigenvalues @xmath397 , the periodic floquet modes @xmath400 suggest the existence of a type of nonlocal invariants within spatially symmetric domains , in analogy to those for stationary states of a static @xmath7 . indeed , integrating the nonlocal continuity [ eq : continuity_nonlocal ] in a floquet state @xmath401 over one period @xmath402 for a potential @xmath403 ( assumed real for simplicity ) which is locally @xmath4-symmetric at any @xmath18 , @xmath404 , we obtain @xmath405_0^\tau = 0\ ] ] since @xmath406 is periodic , where @xmath407 are the period - averaged nlcs in state @xmath401 of the @xmath4-related links @xmath408 . just like in the case of static @xmath7 , we have @xmath409 for equidirectional ( assumed real ) @xmath4-symmetric hoppings @xmath410 , so that the @xmath411 are spatially constant in the domain of symmetry @xmath40 ; cf . [ eq : q_n_+-_transfer ] . this generalizes the characterization of the lss by associated invariants to 1d systems driven periodically with arbitrary strength . like their static counterpart , the period invariants @xmath411 also exist independently of boundary conditions ; they will be illustrated in an example with bound floquet modes in [ sec : inv_per_mod_bound_floquet ] . having established the framework of ls operations , nonlocal current - density continuity and symmetry - induced stationary nonlocal 1d invariants , we now address the occurrence of the latter invariants in bound systems , that is , finite arrays @xmath33 of @xmath8 sites . we first illustrate the various types of ls in 1d , then show how the nlcs relate to symmetry breaking in globally @xmath5-symmetric non - hermitian setups , and finally give an example of nonlocal invariants of bound quasistationary modes in a periodically modulated array . in the remainder of the paper , lengths , field amplitudes and @xmath412-elements are measured in units of @xmath413 ( lattice constant ) , @xmath414 ( amplitude unit ) and @xmath415 ( @xmath412-element unit ) , respectively , which are in turn set to unity , @xmath416 . we consider an array of @xmath417 sites , large enough to encompass different types of symmetry ( global , complete , overlapping , and gapped ; see below ) under local reflection or translation operations . choosing real onsite elements @xmath418 and real equidistant closest - neighbor hoppings @xmath245 , the mode amplitudes @xmath419 of an eigenstate @xmath420 with eigenvalue @xmath421 ( @xmath422 ) can also be chosen real . the nlcs @xmath423 in a general state ( wavepacket ) @xmath424 are composed of a linear combination of the @xmath425 of individual modes @xmath426 as well as mixed - mode currents @xmath427 ( @xmath170 ) which oscillate with frequency @xmath428 . all general statements of [ sec : continuity_nonlocal , sec : stat_invariants_mappings ] about nlcs of general states @xmath56 apply to @xmath429 , while , in particular , the stationary currents @xmath425 are additionally translation invariant in a corresponding ls domain @xmath40 . in the following , we discuss the different types of local @xmath0- and @xmath1-symmetry in terms of the example arrays shown in [ fig : local_symmetries ] together with the locally invariant @xmath425 for selected modes @xmath430 as well as the @xmath307- and @xmath431-matrices producing them via [ eq : q_n_+-_vector ] . + * ( a ) global symmetry . * in the case of global symmetry , there is a connected maximal symmetry domain @xmath432 which extends over the complete setup . we call a ( local ) symmetry maximal @xcite if the corresponding domain @xmath97 is the largest one with the given center @xmath28 ( for @xmath0 symmetry ) or period @xmath29 ( for @xmath1 symmetry ) . in the case of global @xmath0 symmetry , shown in [ fig : local_symmetries](a , i ) , the corresponding permutation matrix @xmath433 is the usual anti - diagonal unit matrix ( upper inset ) , and the @xmath425 vanish for any mode @xmath434 due to its definite parity ( see [ eq : q_p_zero ] ) . if @xmath33 contains a subdomain @xmath435 which is locally @xmath436-symmetric , then its image @xmath437 under the global inversion has opposite invariants @xmath438 , as seen in the lower panel of [ fig : local_symmetries](a , i ) together with the corresponding local inversion matrix ( lower inset ) . we call a finite system ` globally ' @xmath1-symmetric here by convention if it is periodic in @xmath439 , that is , if it consists of repeated copies of a given last period @xmath52 , as shown in [ fig : local_symmetries](a , ii ) ; a ` truly ' globally @xmath1-symmetric array is infinite with @xmath38 . due to the finiteness of @xmath33 , the @xmath425 with @xmath440 differ from the invariants in @xmath40 ( resulting from the inhomogeneity of the global @xmath50-matrix ; see upper inset ) , and therefore we assign the backward invariants @xmath441 ( see [ eq : q_k_-l ] ) to @xmath52 in order to illustrate the @xmath431-symmetry . a locally @xmath442-symmetric subdomain of @xmath33 with different period @xmath443 also leads to different currents , as seen in the lower panel of [ fig : local_symmetries](a , ii ) . + * ( b ) complete local symmetry . * in the case of complete ls ( cls ) , the setup is exactly covered by more than one maximal non - overlapping symmetry domains @xmath444 , that is , @xmath445 with @xmath446 , as exemplified in [ fig : local_symmetries](b ) for local ( i ) @xmath0 and ( ii ) @xmath1 symmetry , excluding here the trivial case of single - site ( size @xmath447 ) domains . the corresponding @xmath49- and @xmath50-matrices are block - diagonal with the blocks being local inversion ( exchange ) and translation ( cyclic shift ) permutation matrices ; see insets . in this class of systems , the local order of the hamiltonian is completely mapped via its eigenstates to the domain invariants @xmath448 which in this sense expose the symmetry information encoded in the states @xmath434 . note that , for both local @xmath0 and @xmath1 symmetry , this information is otherwise not at all visible by the mode amplitude profiles alone for an arbitrary cls setup . considering extended lattice systems of large size @xmath8 , the concept of cls and associated ls - induced invariants may serve to characterize order between the limits of perfect periodicity and uncorrelated disorder , though different from that of quasiperiodicity ( or deterministic aperiodicity ) @xcite . here , the component of disorder may be present in arbitrary amount , within the ls domain themselves or at the level of their distribution in the system . + * ( c ) overlapping local symmetry . * the case of overlapping ls refers to the existence of _ different _ symmetry domains of the same setup with common sites , that is , there are ( composite ) ls domains @xmath449 and @xmath450 with @xmath451 but @xmath452 for some @xmath453 . examples of non - cls setups with overlapping @xmath0- and @xmath1-symmetric domains are shown in [ fig : local_symmetries](c , i ) and ( c , ii ) , respectively . note that overlapping @xmath0-symmetric domains with common center @xmath28 or @xmath1-symmetric domains with common period @xmath454 have equal invariant currents @xmath425 , respectively . in general , a setup may contain mixed local @xmath0 and @xmath1 symmetries in different domains which in turn may overlap ( an obvious case being a @xmath1-symmetric array with @xmath0-symmetric period ) , and the concept of overlapping ls may be used to analyze order in complex systems at different scales @xcite . if overlapping ls domains occur in a cls system , then we speak of different cls _ decompositions _ of the system . as has been shown for continuous models @xcite , scattering setups with multiple complete local @xmath0-symmetry decompositions can enable the design of transparency at multiple incoming frequencies for generic globally @xmath0-asymmetric scatterers . + * ( d ) gapped ( local ) symmetry . * if the hamiltonian elements in one ( or more than one ) subdomain(s ) of a ls domain @xmath40 is modified so that it no longer obeys the considered symmetry ( with a given @xmath0-center or @xmath1-period ) , then the original symmetry domain becomes disconnected , or ` gapped ' , as illustrated in [ fig : local_symmetries](d ) . the @xmath425 are then spatially constant in the symmetric subparts of @xmath40 outside the gap(s ) , but generally unequal in different subparts ( recall that the translational invariance of @xmath425 , pertaining from [ eq : q_n_+-_transfer ] , relies on the connectedness of the ls domain ) . in the case of gapped @xmath0 symmetry , however , the identity ( [ eq : q_p_symmetry ] ) between symmetry - related currents yield equal @xmath425 at symmetry - related subparts for real amplitudes @xmath226 , in spite of the intervening asymmetric gaps . this is seen in [ fig : local_symmetries](d , i ) , where the unconnected symmetry - related left and right ends have equal ( vanishing ) @xmath455 . in the case of gapped @xmath1 symmetry , the gap(s ) induce distinct constant @xmath425 s in each unperturbed part of each gapped period . the example in [ fig : local_symmetries](d , ii ) shows the constancy of @xmath455 in a gapped single - period domain ( the subpart of the first @xmath456 sites of the period @xmath457 is repeated after the gap ) . since the gaps are not symmetric under the considered @xmath4-transformation , the corresponding blocks in the associated permutation matrix are diagonal ( i.e. , applying the identity to that part of the array ) , as seen in the insets of [ fig : local_symmetries](d ) ; note that this yields @xmath458 in the gap in [ fig : local_symmetries](d , ii ) . the gaps may as well be locally symmetric under another transformation ( e.g. @xmath0-symmetric about their center ) , thus yielding a _ nested _ ls , which can be expressed through locally constant @xmath425 s by accordingly modifying those permutation matrix blocks . , * ( b ) * spatially resolved upper nlcs @xmath459 ( offset by site index @xmath47 ) , and * ( c ) * total upper nlc @xmath460 within @xmath40 , in the modes @xmath426 of a * ( i ) * @xmath3-symmetric array with @xmath461 , @xmath462 where @xmath463 and @xmath464 , and a * ( ii ) * @xmath96-symmetric array with @xmath465 , @xmath466 where @xmath467 and @xmath464 , for varying loss / gain parameter @xmath468 . vertical lines indicate exceptional values of @xmath468 where transitions from complex to real eigenvalues @xmath421 occur , corresponding to transitions from complex to ( i ) zero @xmath469 and ( ii ) imaginary @xmath470 in the respective modes . by convention , @xmath471 in ( ii ) ; see text . ] having illustrated the various types of ls in terms of real modes in finite hermitian arrays , we now demonstrate some properties of eigenmode nlcs for minimalistic examples of globally @xmath3- and @xmath96-symmetric non - hermitian arrays , as shown in [ fig : nonherm_modeq_sites ] , and in particular their relation to spontaneous symmetry breaking . the nonhermiticity is here implemented by complex onsite potentials whose imaginary parts @xmath114 represent loss ( @xmath116 ) or gain ( @xmath115 ) of density @xmath117 in time , while the hoppings are taken to be real and equidirectional . global @xmath3 symmetry , for which @xmath472 = 0 $ ] , is known to allow for parameter regimes of loss / gain elements @xmath92 where a subspace of common eigenstates @xmath434 of @xmath7 and @xmath473 with real @xmath7-eigenvalues exists . the rest of @xmath7-eigenstates are ` spontaneously ' symmetry - broken , that is , non-@xmath473-eigenstates , but come in @xmath473-related pairs @xmath474 with complex conjugate eigenvalues @xmath475 . if the whole discrete spectrum is real , then the system is said to be in a @xmath3-unbroken phase . the occurrence of @xmath3-unbroken modes can be seen as originating from the balance between loss and gain in density that a @xmath473-symmetric @xmath7 enables : since @xmath117 in @xmath473 eigenstates is @xmath0-symmetric , the gain on any site @xmath47 is balanced by the loss at site @xmath326 , as seen from [ eq : continuity_local ] , and the current @xmath118 flowing in the state keeps @xmath117 stationary . the @xmath3-unbroken modes are naturally distinguished by vanishing of the corresponding constant nlcs @xmath143 , as follows from [ eq : q_p_zero ] for the case @xmath38 of global symmetry . this is illustrated in [ fig : nonherm_modeq_sites](i ) for a @xmath3-symmetric @xmath476-site array with two loss / gain parameters @xmath477 , where a real central onsite energy @xmath478 is added to break the @xmath479 symmetry of the @xmath7-eigenvalues @xmath480 shown in [ fig : nonherm_modeq_sites](a , i ) . as we see , all @xmath481 vanish within the @xmath468-range ( for fixed @xmath482 ) corresponding to the unbroken phase of real @xmath421 in [ fig : nonherm_modeq_sites](a , i ) . in the @xmath3-broken phase of complex @xmath421 , the @xmath425 are not spatially constant , since @xmath483 for @xmath484 ( and @xmath179 ) in [ eq : continuity_nonlocal_eigenstate ] . note here that @xmath485 and @xmath486 identically in a finite array @xmath33 since they ` flow ' on the nonexisting links @xmath487 and @xmath488 , respectively . while generally @xmath489 in the @xmath3-broken phase , the _ remnant _ of @xmath3 symmetry in the occurrence of @xmath3-related eigenmode pairs is expressed by vanishing sum of nlcs : since @xmath425 in a @xmath7-eigenstate @xmath434 is proportional to the difference of @xmath490 from its @xmath3 image , it will be opposite @xmath491 in the paired eigenstate @xmath492 with @xmath493 , @xmath494 summing separately the @xmath425 along the array , the total upper / lower link nlcs @xmath495 ( here @xmath38 ) in state @xmath434 each serve as a single parameter which distinguish @xmath3-unbroken ( @xmath496 ) from @xmath3-broken ( @xmath497 ) modes for a given setup configuration . this is seen in [ fig : nonherm_modeq_sites](c , i ) for @xmath498 . note that the @xmath499 are imaginary due to [ eq : q_p_symmetry](left ) . further , the remnant of @xmath3 symmetry of the hamiltonian in the broken phase may be expressed by the vanishing sum of @xmath499 over all modes , @xmath500 . the nonlocal charge ( quasipower ) @xmath501 with @xmath502 in the @xmath3-symmetric array for a general state @xmath56 remains constant in time ( longitudinal direction ) @xmath18 , as mentioned in [ sec : nonlocal_charge_evolution ] , in accordance with its vanishing total nlc , @xmath503 . in particular , for the eigenmodes @xmath434 addressed here , we have @xmath504 which shows that the nonlocal charge @xmath505 is zero in @xmath3-broken eigenmodes with @xmath484 . the @xmath3 symmetry of @xmath7 is not a necessary condition for the occurrence of real @xmath421 @xcite and thereby invariant nlcs , of interest here . indeed , real spectra may also occur in @xmath96-symmetric arrays of certain type , which we now exploit to show the induced invariance of translation nlcs @xmath144 . specifically , we implement a non - hermitian version of the su - schrieffer - heeger model @xcite , which can be considered as a special case of the non - hermitian aubry - andr - harper model proposed in ref.@xcite , in an array of odd size @xmath8 where the hoppings are @xmath1-symmetric with period @xmath506 and the potential is simultaneously @xmath96- and @xmath3-symmetric . the latter property enables the presence of loss / gain - balanced eigenmodes with real @xmath421 @xcite . note that such an array is asymmetric under any cyclic shift permutation , that is , under action of an invertible matrix @xmath50 , since it does not consist of an integer number of periods . by convention we thereby set @xmath507 , the largest union of consecutive periods ( starting from @xmath508 ) where @xmath509 , and let @xmath50 be the @xmath506 cyclic shift matrix in [ eq : q_n_+-_vector_k ] . in [ fig : nonherm_modeq_sites](ii ) the behavior of the @xmath470 is shown for the eigenmodes of a @xmath476-site array with a single pair of sites with loss / gain parameter @xmath468 which supports a phase of real @xmath421 in the eigenspectrum shown in [ fig : nonherm_modeq_sites](a , ii ) . like in the @xmath3-symmetric case above , real onsite energies @xmath478 are added to break the @xmath510-symmetry of the spectrum . as we see in [ fig : nonherm_modeq_sites](b , ii ) , the @xmath468-region with real @xmath421 for each eigenmode is distinguished by a spatially constant and imaginary @xmath511 , where @xmath512 is assigned to the last three sites @xmath513 ; see [ eq : q_k_-l ] and below . outside this @xmath468-region , the @xmath514 vary along the array and have a real part , signifying the spontaneous breaking of loss / gain balance in the eigenstates . in analogy to the @xmath3-symmetric case above , though , a remnant of the symmetry of the system can be expressed in the broken phase by the imaginary sums @xmath515 , as anticipated by the @xmath498 shown in [ fig : nonherm_modeq_sites](c , ii ) . the characteristic of imaginary nlcs in the unbroken phase thus survives in the sum of total nlcs within @xmath40 over all modes . sites composed of two domains @xmath435 , @xmath516 which are locally @xmath0-symmetric for the value @xmath517 of a scaling parameter @xmath518 multiplying the onsite potentials @xmath519 ( colored green ) , with periodically modulated hoppings @xmath520 . * ( b ) * time evolution over one period @xmath402 of the nlc in floquet mode @xmath521 associated with the local inversions @xmath436 , @xmath522 . * ( c ) * period averaged nlc for varying scaling parameter @xmath518 , showing the domainwise constancy in the case @xmath517 of local symmetry . ] in [ sec : inv_per_mod ] it was shown how periodically modulated discrete systems reveal local spatial symmetries via the translation invariance of period - averaged nlcs @xmath411 in their quasistationary floquet states . this applies generally for arbitrary boundary conditions and for both @xmath0 and @xmath1 spatial symmetry ( or in combination with @xmath2 ) . the counterpart of the @xmath411 in continuous systems was very recently shown to reliably expose defects in locally @xmath1-symmetric potentials under periodic boundary conditions @xcite . here , the invariance of the period - averaged nlcs in @xmath0-symmetric domains is demonstrated for bound floquet states of discrete setups . in particular , we exploit the discrete tight - binding form of @xmath7 to drive the hoppings in time , corresponding to periodically modulated @xcite waveguide distances in the longitudinal direction of photonic arrays . alternatively , one could drive the onsite potential @xcite , with an equivalent analysis in terms of nlcs . we consider a hermitian array of @xmath523 sites , whose local spatial symmetry is controlled by a scaling parameter @xmath518 which multiplies the real potential @xmath524 on sites @xmath525 , as depicted in [ fig : ls_driven](a ) . for @xmath517 the setup has cls , with the hamiltonian elements of the setup dividing the array into two @xmath456-site @xmath0-symmetric subdomains @xmath435 and @xmath516 with inversion centers @xmath526 and @xmath527 , respectively , while for @xmath528 this ls is broken . although any type of periodic driving is applicable , we here choose to add a simple in - phase harmonic modulation of relative amplitude @xmath529 to all hoppings around the ( initial ) equilibrium hopping strengths @xmath530 , @xmath531 with frequency @xmath532 . for a photonic waveguide array this would correspond to periodically curved waveguides such that the array cross section shrinks and expands homogeneously in the longitudinal direction . the floquet modes @xmath400 of the system are computed with the floquet matrix method @xcite including a total of @xmath456 sidebands ( @xmath533 fourier modes above and below the principal frequency @xmath394 ) . the upper nlc @xmath534 in the floquet state with @xmath521 ( chosen simply to yield distinctly varying nlcs ) are shown in [ fig : ls_driven](b ) for the cls case @xmath517 . as is seen , at any instant @xmath18 we have that @xmath535^$ ] in each @xmath0-symmetric subdomain from [ eq : q_p_symmetry ] ( left ) which clearly holds also for the present dynamical nlcs . what is not directly anticipated from the time evolution of the nlcs is that their averages @xmath411 over a period @xmath402 are real and spatially constant within @xmath435 and @xmath516 , as was established in [ sec : inv_per_mod ] . the profile of the period - averaged nlcs in relation to ls is illustrated in [ fig : ls_driven](c ) by continuously varying @xmath518 . for @xmath536 the system has no symmetry around @xmath537 in @xmath435 , and the @xmath538 are symmetric on the links around @xmath537 . the subdomain @xmath516 , where only one site is modified by @xmath518 , has a gapped @xmath0 symmetry , and the @xmath539 are spatially constant in the subdomain @xmath540 ( i.e. , only @xmath541 on the central link across @xmath542 deviated from the rest ) . in both domains we see that @xmath543 becomes spatially constant in the limit of cls when @xmath517 . alternatively to the bound array considered here , an interesting perspective would be to study the behavior of nlcs in driven scattering setups where , e.g. floquet bound states @xcite may arise in the energy continuum . in the previous section the nlcs and their invariance in ls domains , were investigated in various types of bound setups with discrete eigenspectra . we now apply the framework of nlcs to stationary _ scattering _ , where the energy @xmath20 ( longitudinal propagation constant for photonic waveguides ) is a given input parameter , to demonstrate their domainwise invariance and relation to transmission in discrete structures with lss . to simulate scattering of an incident monochromatic wave on a localized potential in the discrete system , we attach semiinfinite chains with real constant onsite elements @xmath232 and hoppings @xmath544 on the left ( @xmath545 ) and right ( @xmath546 ) of the @xmath8-site scatterer . in the homogeneous chain leads , stationary solutions of the se ( [ eq : schr ] ) are linear combinations of plain wave states @xmath547 where @xmath548 with quasimomentum @xmath75 for energies obeying the dispersion relation @xmath549 for ` biased ' setups with different lead hamiltonians ( @xmath550 and/or @xmath551 ) the plain wave states are flux normalized by the respective factor @xmath552^{1/2}$ ] with @xmath75 given by the lead dispersion ; we here consider unbiased setups with equal @xmath553 in the two leads for simplicity . the spatial amplitudes @xmath226 of a scattering state @xmath554 are expanded in the plain waves of the leads as ( @xmath555 ) @xmath556 on the left and right of the scatterer , respectively . the computation of the amplitudes @xmath226 along the scattering region as well as the output amplitudes @xmath557 for given input is presented in compact form in [ app : scattering ] . an incoming plain wave with unit amplitude from the left , setting @xmath558 ( from the right , setting @xmath559 ) , is reflected with amplitude @xmath560 ( @xmath561 ) and transmitted with amplitude @xmath562 ( @xmath563 ) . these scattering amplitudes define the scattering matrix @xmath564 which maps general input amplitudes to their output in the @xmath565 basis , @xmath566 the reflection and transmission coefficients are given by @xmath567 and @xmath568 , respectively , with @xmath569 dictated by current conservation in the absence of ( imbalanced ) loss and gain . in the following , we will focus on ( local ) @xmath0-symmetric scatterers , whose corresponding nlcs are to be related to perfect transmission ( @xmath570 ) of composite setups . the nlc on the first links on the left and right of a scatterer localized in a domain @xmath40 can generally be written in terms of the in- and outgoing amplitudes in the @xmath565 basis as @xmath571 & = ih ( \zeta^ - \zeta ) \begin{bmatrix}\ , { { { \mathfrak{c}}}}_<^+ \ , \vspace{.7ex } \\ { { { \mathfrak{c}}}}_<^- \end{bmatrix}^\dagger \mathsf{a}^\dagger \mathsf{\lambda } \ , \mathsf{a } \begin{bmatrix}\ , { { { \mathfrak{c}}}}_>^+ \ , \vspace{.7ex } \\ { { { \mathfrak{c}}}}_>^- \end{bmatrix}\end{aligned}\ ] ] where @xmath572 -1 & 0 ~\end{bmatrix } } $ ] , @xmath573 correspond to a shift of the scatterer by @xmath574 to the origin via the matrix @xmath575 $ ] , with the center @xmath28 of @xmath40 taken to coincide with the center of inversion ( we suppress the subscript @xmath0 in @xmath576 from here on ) . note that , since the scattering matrix @xmath564 connects @xmath565-amplitudes , the above expression for the nlc remains essentially the same for a continuous system : the discreteness of the system enters merely through the factor @xmath577 , which is replaced by @xmath578 in the continuum limit ( with real @xmath579 ) where the nlc becomes a two - point quantity @xmath580 . substituting the outgoing amplitudes @xmath581 from [ eq : smatrix ] , the nlc above takes the form @xmath582 where @xmath583 , @xmath584 and @xmath585 are the ingoing and reflection amplitudes , respectively , for the same scatterer shifted to the origin ( @xmath586 ) , and where it was used that @xmath587 in general @xcite for nonsingular hamiltonians . -symmetric hermitian array composed of three ls domains : @xmath588 and @xmath589 each containing a scatterer with hamiltonian parameter @xmath590 ( top sketch ) , and an intervening free propagation domain @xmath591 . reflection magnitudes * ( a1 ) * @xmath592 of @xmath588 , * ( a2 ) * @xmath593 of @xmath589 , and * ( a3 ) * @xmath594 of the composite array @xmath595 , as well as quantities * ( a4 ) * @xmath596 , * ( a5 ) * @xmath597 , and * ( a6 ) * @xmath598 , which vanish at unit transmission ( ptr , @xmath599 ) , for varying @xmath590 and @xmath75 . an sptr ( aptr ) occurs at @xmath600 where @xmath601 ( @xmath602 ) , in which case @xmath603 ( @xmath602 ) , indicated by @xmath604 ( @xmath605{45}{$\square$}}}$ ] ) . * ( b1 ) * transmission @xmath606 and nlcs @xmath607 ( offset by @xmath608 and scaled by overall maximum ) in varying @xmath75 for @xmath609 and * ( b2 ) * density @xmath82 ( offset by @xmath610 ) at @xmath75-values indicated by @xmath611{60}{$\triangle$}}$ ] in zoomed inset of ( b1 ) where @xmath612 ( corresponding to the sptr in ( a ) ) , @xmath613 , @xmath614 , respectively . @xmath82 is @xmath615-symmetric when @xmath616 ( @xmath617 ) . * ( c1 ) * @xmath606 and @xmath618 for @xmath619 and * ( c2 ) * @xmath82 at @xmath75 indicated by @xmath605{45}{$\square$}}}$ ] in inset where @xmath620 corresponding to the aptr in ( a ) ; green dashed line shows @xmath621/\pi$ ] . @xmath588 ( @xmath589 ) is here augmented symmetrically by 1 ( 2 ) site(s ) into @xmath622 ( @xmath623 ) . ] let us now assume the scatterer to be @xmath0-symmetric within a connected domain @xmath40 and hermitian with equidirectional hoppings , meaning that @xmath624 is translation invariant along @xmath40 . we then also have that @xmath625 from the @xmath0 symmetry of the scatterer ( around the origin ) and @xmath626 , @xmath627 from the @xmath564-matrix unitarity , so that the nlc expression reduces to @xmath628 where we note that @xmath629 . this expression shows how the @xmath630 depend on the characteristics @xmath631 , @xmath632 and position @xmath28 of the scatterer ( domain @xmath40 ) as well as the amplitudes incident on it . at the center @xmath633 and edges @xmath634 of the leads brillouin zone , the scattering nlcs vanish generically . away from those @xmath75-points , @xmath635 and thereby symmetric density @xmath117 in @xmath40 ( see [ eq : q_p_zero ] ) can be achieved by tuning the ingoing amplitudes to satisfy @xmath636 = \pm 2 \|{{\mathfrak{t}}}\|/\|{{\mathfrak{r}}}\|$ ] for @xmath637 . for example , this is satisfied for @xmath638 and @xmath639 at all @xmath20 where @xmath640 . a special situation occurs for @xmath564-matrix eigenstates , in which @xmath641 and thus @xmath635 for any @xmath642 . for a globally @xmath0-symmetric ( about @xmath643 ) hermitian scatterer , those states are simultaneously @xmath644 eigenstates with eigenvalues @xmath645 and @xmath473 eigenstates with eigenvalues @xmath646 where @xmath647 ( note that @xmath648 in this case ) . in the most common setting for a single scatterer , however , there is only one incoming wave from the left ( @xmath649 ) or from the right ( @xmath650 ) . in this case , @xmath635 if and only if the scatterer is transparent at the given @xmath75 , @xmath599 ( or if @xmath651 , which strictly only occurs for singular potentials in unbiased 1d systems , in contrast to fano transmission zeros when multiple interference paths are available @xcite ) . thus , perfect transmission @xmath652 is accompanied by a @xmath307-symmetric density @xmath117 within @xmath40 ( @xmath306-symmetric @xmath653 up to a phase ) . the assumption @xmath654 for left ( right ) incidence is compatible with the connection to another , generally different , locally symmetric scatterer on the right ( left ) of @xmath40 which is also transparent at the same @xmath75 . by adjusting the hamiltonian parameters , a cls setups may be assembled from multiple , simultaneously transparent scatterers in ls domains @xmath66 ( @xmath655 ) , with the state density following the ls of the setup . we refer to such a state as a ` symmetric ' perfect transmission resonance ( sptr ) ; an example is illustrated in [ fig : ptr ] . we here have two scatterers localized in domains @xmath588 and @xmath589 , connected via a free ( with @xmath656 like in the leads ) gap @xmath591 . as we see in [ fig : ptr](b ) , the vanishing of the individual @xmath657 reveals the @xmath75 at which the density @xmath117 is locally symmetric in @xmath658 , respectively , and an sptr occurs when @xmath659 . in contrast to an sptr , transparency ( @xmath652 ) may also occur in a state whose density does not follow any ls of the setup , which we refer to as an ` asymmetric ' ptr ( aptr ) . to see the relation of the nlcs to the occurrence of an aptr , we consider a hermitian setup of two different scatterers @xmath435 and @xmath516 with corresponding scattering matrices @xmath660 and @xmath661 . the total @xmath564-matrix has elements @xmath662 , @xmath663 , and @xmath664 , where @xmath665 accounts for all orders of multiple reflections between the scatterers . the condition @xmath666 for a ptr leads to @xmath667 , with @xmath668 being the phases of @xmath669 , which in turn is equivalent to @xmath670 as can be seen geometrically in the complex plane . thus , in the special case @xmath671 , the phases @xmath672 are arbitrary , while @xmath673 must hold for @xmath674 @xcite . if both scatterers are @xmath0-symmetric about their centers , it additionally holds that @xmath675 ( @xmath676 ) , and it can then be shown that @xmath677 at the ptr . that is , the phase of the transmission amplitude depends ( up to @xmath678 ) only on the distance between the centers of _ any _ two arbitrary but locally @xmath0-symmetric scatterers . assuming the scatterers to be @xmath0-symmetric , each will also have spatially constant nlc @xmath679 ( @xmath680 ) at a given @xmath20 . using the general expression ( [ eq : q_scat_in - out - amps ] ) , the nlcs in @xmath435 can be written as @xmath681 in terms of the global in- and outgoing @xmath565-amplitudes @xmath682 , where the ` connection ' matrix @xmath683 was used to eliminate the amplitudes between the scatterers ( note that the @xmath435 and @xmath516 may be directly attached in the above formulation ) . considering now the case of a ptr ( @xmath599 , @xmath684 ) under left incidence only ( @xmath685 ) , the upper nlc becomes @xmath686 on the other hand , its difference from @xmath687 is generally given by @xmath688 for left incidence . in the case of a ptr , applying conditions ( [ eq : ptr_condition ] ) and ( [ eq : t_ptr ] ) makes [ eq : q_scat_d1_d2_diff ] consistent with [ eq : q_scat_d1 ] only if @xmath689 for @xmath690 , the sign being determined by the details of a given setup . this is demonstrated in the example setup of [ fig : ptr](c ) , featuring an aptr where the nonzero nlcs @xmath691 and @xmath692 cross . the relation between nlcs and transparency of a two - scatterer setup can be verified for a setup with a gap of two ` free ' sites @xmath693 ( @xmath694 , @xmath695 ) between @xmath435 and @xmath516 , for which an explicit calculation yields @xmath696 for @xmath666 . since the @xmath697 , @xmath698 are constant and imaginary , we have that @xmath699 , in accordance with [ eq : t_ptr ] . to summarize , when the setup of two ls scatterers is perfectly transmitting , then it always holds that @xmath700 , with two possible cases with respect to the symmetry of @xmath117 : ( i ) if @xmath701 , then we have an sptr with @xmath117 being @xmath0-symmetric in the ls domains . ( ii ) if @xmath702 , then we have an aptr and @xmath117 is asymmetric with respect to the ls domains . the nlcs thus connect the transport properties of globally asymmetric systems to the spatially resolved profile of the scattering state amplitude with respect to its possible lss . in composite cls setup of multiple scatterers , there may be parts of locally symmetric and asymmetric @xmath117 at a ptr indicated by zero and nonzero equal nlc invariants , respectively . -symmetric non - hermitian array @xmath703 with onsite elements @xmath704 ( @xmath705 ) and common hoppings @xmath706 . * ( a1 ) * quantity @xmath707 of @xmath40 and nlcs * ( a2 ) * @xmath708 of @xmath588 and * ( a3 ) * @xmath709 of @xmath589 for unit incoming amplitude from the left , in varying @xmath75 and @xmath710 for fixed @xmath711 . the colormap is normalized to @xmath608 in ( a1 ) and truncated at @xmath712 in ( a2,a3 ) ; it scales with the power of @xmath713 to increase contrast . * ( b ) * density @xmath82 ( @xmath714 ) and spatially resolved nlcs @xmath715 ( @xmath604 ) for local @xmath716 transforms , @xmath717 ( @xmath718 ) and @xmath719 ( @xmath720 ) for global @xmath0 transform , for @xmath721 ( global @xmath3 symmetry ) at @xmath722 indicated by upper dotted line in ( a1 ) . * ( c ) * same as ( b ) but for @xmath723 ( global @xmath0 symmetry ) at @xmath724 indicated by lower dotted line in ( a1 ) . ] we close this section by addressing the nlcs of scattering states in non - hermitian setups . since the energy is here a free input parameter which is chosen real , the translation invariance of the nlcs in locally @xmath3-symmetric domains is manifest for any choice of loss / gain parameters . in the case of a @xmath3-symmetric scatterer , the @xmath564 matrix element are subject to the generalized unitarity relation @xcite @xmath725 for @xmath726 , with @xmath727 . the general expression ( [ eq : q_scat_in_t_r_general ] ) for the nlc then becomes @xmath728 which further reduces to [ eq : q_scat_in_t_r_herm_psymm ] if the @xmath3-symmetric scatterer is also hermitian . a @xmath3-unbroken phase for a globally @xmath3-symmetric , two - terminal scattering setup may be identified as the parametric region of @xmath7 with unimodular @xmath564-matrix eigenvalues @xmath729 at given energy @xcite . indeed , for @xmath564-matrix eigenstates , where @xmath730 , the nlc of [ eq : q_scat_in - out - amps ] becomes @xmath731 yielding a @xmath473 eigenstate when @xmath732 since then @xmath733 . the latter condition , however , can be met in [ eq : q_scat_in_t_r_ptsymm ] also for non-@xmath564-matrix eigenstates in dependence of the incoming amplitudes @xmath734 for a given @xmath735 . the crossing from zero to nonzero nlc thus provides an extended identification of a @xmath3 ` phase transition ' in scattering beyond @xmath564-matrix eigenstates @xcite , identical to that of bound systems ( as discussed above in [ sec : non - hermitian ] ) . more importantly , the vanishing of nlcs can be employed to identify configurations with _ locally _ @xmath3-invariant states , that is , where [ eq : q_p_zero ] is fulfilled in a @xmath3-symmetric subdomain of the setup . as an example , in [ fig : local_pt ] we consider a non - hermitian setup of two attached domains @xmath588 , @xmath589 with loss / gain parameter pairs @xmath736 , @xmath737 , respectively . the setup is designed such that it is globally @xmath3-symmetric ( @xmath0-symmetric ) for @xmath738 ( @xmath739 ) . as we see in [ fig : local_pt](a1 ) , there are values of @xmath740 where [ eq : conservation_pt ] is fulfilled at certain energies ( see white contours ) , such that the system behaves as being @xmath3-symmetric in terms of its @xmath564-matrix . interestingly , it is possible to fulfill [ eq : conservation_pt ] at all @xmath75 with a locally @xmath3-symmetric setup for appropriately chosen loss / gain parameters , as seen for the present setup for @xmath741 ( lower horizontal white line in [ fig : local_pt](a1 ) ) . in [ fig : local_pt](a2 ) and ( a3 ) we see that the nlcs @xmath742 and @xmath743 , corresponding to local @xmath3-symmetry in @xmath588 and @xmath589 , vanish at different contours in the @xmath744-plane , revealing the occurrence of locally symmetric density embedded into the globally @xmath3-asymmetric system . leaving a more comprehensive investigation of nlcs in relation to scattering properties of locally @xmath3-symmetric setups for future work , we focus now on their spatial constancy in 1d domains . this is illustrated in [ fig : local_pt](b ) and ( c ) , where the considered setup is , apart from locally @xmath3-symmetric , also globally @xmath3- and @xmath0-symmetric , respectively . at the chosen generic @xmath75-values , the density profile @xmath82 is completely asymmetric under any of those symmetry transformations at the chosen ( generic ) @xmath75-values . for the local @xmath0 transforms in @xmath588 and @xmath589 , the domainwise constancy of the associated nlc @xmath745 reveals the local @xmath3 symmetry of @xmath7 encoded in the scattering state . note that this is achievable because @xmath20 is set real as an input parameter , in contrast to non - hermitian _ bound _ 1d systems where local @xmath3 symmetry typically does not support real eigenenergies . in [ fig : local_pt](b ) , the constant nlc @xmath746 associated with global @xmath0 reflects the global @xmath3 symmetry , while the alternative quantity @xmath747 is symmetric ; see [ eq : q_p_symmetry ] . in [ fig : local_pt](c ) , we have the opposite situation : @xmath746 is symmetric , while the constant @xmath747 reflects the global @xmath0 symmetry , even in the presence of loss and gain . a discrete nonlocal current - density continuity formalism was developed to relate localized spatial symmetries to schrdinger wave properties in hermitian and non - hermitian lattice systems . sources and sinks of the nonlocal currents ( nlcs ) are identified in the breaking of the local symmetries ( lss ) , with the associated charge evolution governed by the nlcs at symmetry domain boundaries . for stationary states , the symmetry - adapted nlcs are translationally invariant in one - dimensional finite domains with @xmath273-symmetric sub - hamiltonians , where @xmath6 ( inversion ) or @xmath1 ( translation ) is potentially combined with @xmath2 ( time reversal ) . these invariant nlcs reflect the local symmetries of generic single - particle lattice hamiltonians encoded into arbitrarily irregular field profiles , and enable the mapping between wave amplitudes of symmetry - related sites in finite domains . we illustrated the nlc invariance in cases of complete , overlapping , and gapped domainwise symmetry , as well as for non - hermitian ( locally ) @xmath3- and @xmath96-symmetric ( scattering ) systems with balanced gain and loss . in scattering setups , the nlcs were further shown to classify perfect transmission resonances with respect to their spatially resolved density profile , being symmetric ( asymmetric ) in adjacent locally symmetric lattice subdomains with equal and zero ( nonzero ) nlc magnitude . further , in periodically driven setups the nlc invariance was shown to be retained for quasi - stationary ( floquet ) states on period - average , thus revealing lss also far from equilibrium . we have here utilized photonic waveguide arrays as an applicational platform since they are well described by a discrete schrdinger equation with nearest neighbor couplings . the formalism and general results apply equally to other lattice systems modeled effectively by a discrete schrdinger equation , like superlattice nanostructures or quantum dot arrays @xcite . for stationary states the applicability is further broadened to generic wave mechanical systems described by a discretized helmholtz equation , such as e.g. granular acoustic media with domainwise homogeneities @xcite . in the limit of large site density with constant inter - site hopping , the same formalism can be equally employed for spatially continuous systems ( with the lattice used as a numerical grid ) , treating simulations of , e.g. , photonic multilayer devices @xcite or acoustic waveguides @xcite . the present operator formulation of the nlcs , expressed in [ eq : q_n_+-_p_operator ] , further enables a straightforward extension beyond nearest neighbor couplings or to higher dimensions , simply by including the additional couplings into the hopping operators @xmath124 accordingly . in general , the net nlc assigned to a site with ( multidimensional ) index @xmath748 will then consist of the sum of nonlocal link currents @xmath749 over connected sites @xmath750 for a given symmetry operation @xmath4 , defined in analogy with the 1d nlcs in [ eq : q_n+- ] . a thorough account on symmetry considerations and associated nlcs in planar discrete systems , where rotations and plane reflections add to the possible @xmath4-transforms , will be presented in a future work @xcite . multidimensional divergence naturally prevents individual @xmath751 from being spatially constant along a selected dimension , just as is the case for the usual current . nlc invariance does apply if the sites @xmath750 above represent next - nearest or ( selected ) remote neighbors along a _ dimension , as is effectively the case , e.g. , for zigzag chains with embedded scatterers @xcite or discrete helical structures @xcite , respectively . extended locally symmetric domains in such setups will then be distinguished by constant nlcs in arbitrary stationary states . finally , the domainwise invariant stationary nlcs may be generalized to discrete interacting systems where the _ interaction _ itself is represented by a locally @xmath5-symmetric function on the lattice sites . a promising case would be the discrete version of self - induced @xmath3-symmetric potentials @xcite in lattice subdomains , with the mean - field nonlinear ( gross - pitaevskii ) term @xmath752 replaced by the nonlocal density @xmath753 . with the above perspectives , we believe that the developed connection between discrete nonlocal continuity and symmetry - induced subdomain invariants may contribute to the understanding of wave structure and response in discrete systems with localized spatial regularities . we are thankful to malte rntgen for fruitful suggestions and critical reading of the manuscript , as well as to steffen weimann and alexandra v. zampetaki for helpful discussions . augmenting the scatterer domain @xmath33 by two lead sites on each side ( @xmath754 and @xmath755 in the present indexing ) , we can solve the close - coupling equations ( [ eq : schr_cc ] ) for the outgoing plane wave amplitudes @xmath581 as well as the site amplitudes @xmath756^\top\ ] ] along the scatterer by sparse matrix multiplication and inversion of size @xmath757 . specifically , using the asymptotic ansatz ( [ eq : scat_bc ] ) for the two sites on either side of @xmath33 , we rearrange the schrdinger system of @xmath757 next - neighbor equations to solve for @xmath581 and @xmath163 in the form of the column vector @xmath758^{-1 } \begin{bmatrix}\ , ( h^\zeta^ + \tilde{v } ) \zeta^{+l}\ , { { \mathfrak{c}}}_<^+ \ , \\ { { \boldsymbol{\zeta}}}^ < _ > \\ ( h^\zeta^ + \tilde{v } ) \zeta^{-r } \ , { { \mathfrak{c}}}_>^- \,\end{bmatrix}.\end{aligned}\ ] ] here , @xmath759 ( @xmath760 ) is the index of the lead site attached to the leftmost ( rightmost ) scatterer site , @xmath761 , and @xmath762 with @xmath763 being a @xmath764 row vector of zeros and @xmath765 the @xmath162 unit matrix . the @xmath766 matrix @xmath767 is the @xmath768 block of @xmath769 corresponding to the domain @xmath770 , augmented by the column @xmath771^\top$ ] on the left and @xmath772^\top$ ] on the right . with the scatterer domain defined as @xmath773 we here have @xmath774 and @xmath775 . in the case @xmath776 of a single site scatterer , @xmath777 reduces to a single element @xmath778 note also that for biased setups , the @xmath779 and @xmath75 entering the top ( bottom ) row elements in [ eq : scat_general_output ] are accordingly replaced by those of the left ( right ) lead ; in the present work we consider unbiased setups . the above method is essentially equivalent to propagating the input signal via the resolvent ( lattice green function ) of the effective lead - coupled hamiltonian of the scattering region . here , however , we explicitly solve outgoing @xmath565-amplitudes and combine them compactly into a column vector with the spatially resolved field in the scatterer domain @xmath33 . this formulation of the scattering problem has an advantage over a standard site - by - site transfer matrix method in that the scatterer can be of arbitrary type : keeping a two - terminal geometry , the scatterer may contain , e.g. , side - coupled sites or a 2d cluster of interconnected sites . no explicit single - node transfer matrices propagating the wave along the scatterer need to be specified , rather the complete hamiltonian under appropriate indexing is utilized directly . + the nlcs @xmath169 and their alternatives @xmath264 for a general state @xmath56 can be expressed together via a matrix product acting on a column vector of amplitudes in @xmath780 , as @xmath781 for arbitrary hamiltonian , with @xmath170 . for nonzero local currents @xmath122 , the first matrix above ( with determinant @xmath782 ) is invertible and we can solve for the symmetry - transformed amplitudes to obtain @xmath783 the lower equation above yields the mapping relation in [ eq : current_mapping ] from @xmath784 to @xmath331 , where the upper ( @xmath167 ) _ or _ lower ( @xmath168 ) nlcs can be equally used . the upper equation yields the mapping relation between the sites adjacent to @xmath338 using the _ same _ ( correspondingly upper or lower ) nlcs , provided that the involved hoppings are real and @xmath4-symmetric , @xmath785 . multiplying [ eq : current_mapping_matrix_general ] from the left by the row vector @xmath786 \nonumber\ ] ] we arrive the current identity in [ eq : current_connection ] . using the bitemporal nlc @xmath192 instead of @xmath264 would also lead to a mapping relation analogous to [ eq : current_mapping ] , though mapping @xmath787 to @xmath788 and involving accordingly a bitemporal local current @xmath789 in the case of a stationary state @xmath236 ( with real @xmath20 ) , the stationary mapping relation of [ eq : current_mapping_stat ] is identically recovered , since the temporal factors @xmath790 are factored out . 71ifxundefined [ 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 _ _ , ed . ( , , ) link:\doibase 10.1103/revmodphys.82.3045 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.235126 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.114.245503 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.67.516 [ * * , ( ) ] link:\doibase 10.1063/1.1536639 [ * * , ( ) ] link:\doibase 10.1103/physreva.75.063424 [ * * , ( ) ] link:\doibase 10.1088/0034 - 4885/69/2/r03 [ * * , ( ) ] link:\doibase 10.1016/s0370 - 1573(02)00559 - 8 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.013901 [ * * , ( ) ] link:\doibase 10.1007/s11071 - 014 - 1422 - 1 [ * * , ( ) ] link:\doibase 10.1073/pnas.0905337106 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.100.013906 [ * * , ( ) ] link:\doibase 10.1088/0953 - 8984/25/10/105401 [ * * , ( ) ] link:\doibase 10.1021/jp011482j [ * * , ( ) ] link:\doibase 10.1107/s0021889808033384 [ * * , ( ) ] http://ebooks.cambridge.org/ebook.jsf?bid=cbo9780511626128[__ ] ( , , ) link:\doibase 10.1103/physreva.81.053808 [ * * , ( ) ] link:\doibase 10.1063/1.1468895 [ * * , ( ) ] link:\doibase 10.1063/1.4801890 [ * * , ( ) ] link:\doibase 10.1088/1367 - 2630/16/9/093017 [ * * , ( ) ] link:\doibase 10.1364/josab.28.002584 [ * * , ( ) ] link:\doibase 10.1103/physreva.87.032113 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.050403 [ * * , ( ) ] link:\doibase 10.1088/1751 - 8113/49/19/195304 [ * * , ( ) ] link:\doibase 10.1103/physreva.88.033857 [ * * , ( ) ] link:\doibase 10.1103/physreva.90.043809 [ * * , ( ) ] link:\doibase 10.1103/physreva.93.063831 [ * * , ( ) ] link:\doibase 10.1103/physrevb.92.014303 [ * * , ( ) ] http://www.worldscientific.com/doi/abs/10.1142/s0217732301005333 [ * * , ( ) ] link:\doibase 10.1088/0305 - 4470/35/7/315 [ * * , ( ) ] link:\doibase 10.1088/1751 - 8113/48/15/155304 [ * * , ( ) ] link:\doibase 10.1098/rsta.2012.0194 [ * * , ( ) ] link:\doibase 10.1103/physreve.91.042908 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.80.5243 [ * * , ( ) ] \doibase 10.1023/b : cjop.0000044016.95629.a7 [ * * , ( ) ] link:\doibase 10.1016/j.aop.2006.05.011 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.093902 [ * * , ( ) ] link:\doibase 10.1103/physrevx.3.041030 [ * * , ( ) ] link:\doibase 10.1103/physreva.92.022125 [ * * , ( ) ] link:\doibase 10.1103/physreva.93.022102 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.85.1863 [ * * , ( ) ] link:\doibase 10.1103/physreva.90.053817 [ * * , ( ) ] link:\doibase 10.1103/physreva.92.041803 [ * * , ( ) ] link:\doibase 10.1038/ncomms8782 [ * * , ( ) ] link:\doibase 10.1103/physrevb.44.10637 [ * * , ( ) ] link:\doibase 10.1088/0143 - 0807/31/5/010 [ * * , ( ) ] link:\doibase 10.1364/oe.15.001579 [ * * , ( ) ] in http://link.springer.com/chapter/10.1007/978-3-662-04506-0_2[__ ] ( , ) , pp . link:\doibase 10.1103/physrevlett.113.123903 [ * * , ( ) ] link:\doibase 10.1103/physreve.89.052918 [ * * , ( ) ] @noop _ _ ( , , ) link:\doibase 10.1103/physreva.78.062113 [ * * , ( ) ] link:\doibase 10.1103/physreva.89.062102 [ * * , ( ) ] link:\doibase 10.1103/physreva.93.062101 [ * * , ( ) ] link:\doibase 10.1103/physreve.93.052215 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.101.203902 [ * * , ( ) ] link:\doibase 10.1016/j.physrep.2012.03.005 [ * * , ( ) ] link:\doibase 10.1103/physrevb.73.193305 [ * * , ( ) ] link:\doibase 10.1016/s0370 - 1573(98)00022 - 2 [ * * , ( ) ] link:\doibase 10.1038/srep02219 [ * * , ( ) ] @noop _ _ ( , ) link:\doibase 10.1103/physreva.85.023802 [ * * , ( ) ] link:\doibase 10.1103/physreve.71.036626 [ * * , ( ) ] link:\doibase 10.1063/1.3556455 [ * * , ( ) ] link:\doibase 10.1103/physreva.92.023605 [ * * , ( ) ] link:\doibase 10.1103/physreva.91.023606 [ * * , ( ) ]	a nonlocal discrete continuity formalism is developed which relates spatial symmetries in subparts of hermitian or non - hermitian lattice systems to the properties of adapted nonlocal currents . broken local symmetries thereby act as current sources or sinks , and the time evolution of the associated nonlocal charge is governed by the nonlocal currents at the boundaries of domains with local symmetry . we apply the framework to locally inversion-(time- ) and translation-(time- ) symmetric one - dimensional photonic waveguide arrays effectively described by schrdinger s equation with a tight - binding hamiltonian . the nonlocal currents of stationary states are shown to be translationally invariant within local symmetry domains for arbitrary wavefunction profiles , and cases of complete , overlapping , and gapped local symmetry are demonstrated for model setups . two distinct versions of the nonlocal invariant currents enable a mapping between wave amplitudes of symmetry - related sites , thereby generalizing the global bloch and parity mapping to local translation and inversion symmetry in discrete systems . in scattering setups , perfectly transmitting states are characterized by aligned nonlocal invariants in attached symmetry domains , whose vanishing signifies a correspondingly symmetric density . for periodically driven arrays , the invariance of the nonlocal currents is retained on period average for quasi - energy eigenstates . encoding local potential and coupling symmetries into arbitrary stationary states , the theory of symmetry - induced continuity and local invariants may contribute to the understanding of wave structure and response in systems with localized spatial order .
it was a remarkable discovery of the beginning of this century , anticipated in physics in the nineties , that the degree zero small quantum cohomology of a smooth deligne mumford stack @xmath4 does not in general coincide with the ordinary cup product , and that it is , in fact , a proper ring extension of the ordinary cohomology ring of @xmath4 . more generally , the definition of quantum cohomology and gromov witten invariants for orbifolds were given in symplectic geometry by chen and ruan in @xcite . the algebraic counterparts of these theories were developed by abramovich vistoli in @xcite , @xcite . ruan cohomology of a smooth deligne mumford stack @xmath4 is , by definition , the degree zero part of the small quantum cohomology ring of @xmath4 , and the orbifold cohomology of @xmath4 is the rationally graded vector space that underlies the chen ruan cohomology algebra . the general idea , coming from stringy geometry , is that an important role in the study of @xmath4 is played by the so - called _ inertia stack _ of @xmath4 . when @xmath4 is a moduli space for certain geometric objects , the inertia stack of @xmath4 parametrizes the same geometric objects , together with the choice of an automorphism on them . the stack @xmath4 itself appears as the connected component of its inertia stack associated with the trivial automorphism , but in general there are other connected components , usually called the _ twisted sectors _ of @xmath4 , a terminology that originates from physics . orbifold cohomology is simply the ordinary cohomology of the inertia stack , endowed with a different grading . each twisted sector is assigned a rational number , called ( depending on the author ) _ degree shifting number _ , _ age _ or _ fermionic shift _ : this number depends on the action of the given automorphism on the normal bundle to the twisted sector in @xmath4 . then the degree of each cohomology class of the twisted sector is shifted by twice this rational number . in this paper , we study the inertia stack of moduli spaces @xmath5 of smooth genus @xmath0 curves . the starting point of our construction is that one can associate with each object @xmath6 of the inertia stack the cover given by quotienting @xmath7 by the cyclic group generated by @xmath8 . following an idea of fantechi @xcite , we exploit this correspondence to tackle the problem of the identification of the twisted sectors of @xmath5 by using the classical theory of cyclic ( possibly ramified ) covers of algebraic varieties , as developed in @xcite . we identify some discrete data in order to separate the inertia stack of @xmath9 in its connected components . the first data are the genus of the quotient curve and the order @xmath10 of the automorphism ; the latter is a general invariant of twisted sectors as it appears already in the definition of the inertia stack . finally , the branch locus of the cover can be split in @xmath11 parts according to the local monodromy around each of its points . the last invariants are simply the degrees of each of these @xmath11 divisors . it is a recent result of catanese @xcite that these numerical data single out a connected component of the moduli space of connected cyclic covers . the twisted sectors of moduli of curves were studied with _ ad hoc _ methods in the case of genus @xmath12 in @xcite , and in the case of pointed curves of genus @xmath13 in @xcite . the same approach explained in the above paragraph was used in @xcite to identify the twisted sectors of @xmath14 with @xmath15 or @xmath16 , in this paper we complete the picture by analyzing the more delicate case when @xmath17 . a complete cohomological description of the twisted sectors of moduli of hyperelliptic curves for all genera is given in @xcite following a similar approach . after having determined the connected components of the inertia stack of @xmath5 for general @xmath0 , we study the topology of the twisted sectors in the case when @xmath0 equals @xmath2 . in most cases , the cohomology of the twisted sector is computed in a rather straightforward way . the main exceptions are the twisted sectors corresponding to bielliptic and to quadrielliptic genus @xmath2 curves , which require a more detailed analysis . in particular , our computation of the cohomology of the moduli space of bielliptic genus @xmath2 curves is achieved by using a combination of vassiliev gorinov s method for the computation of the cohomology of complements of discriminants with the study of certain leray spectral sequences , following the approach of @xcite , @xcite . we expect that these techniques could be applied also in other cases of moduli spaces of cyclic covers , at least for small values of @xmath0 . finally , we partially extend our investigation to the orbifold cohomology of the deligne mumford compactification @xmath18 of @xmath3 . specifically , we study the zariski closure of @xmath19 inside the inertia stack @xmath20 . the connected components of this compactification are precisely the connected components of @xmath21 whose general element is a smooth curve . the study of the compactification of the inertia stack is performed using moduli stacks of admissible covers ( we refer for the general theory to @xcite ) . the cohomology of the twisted sectors contained in this compactification contributes to what we call the compactified orbifold cohomology of @xmath3 . the main results of the paper are the description of the connected components of the twisted sectors of @xmath9 ( sect . [ sectiondefinertia ] ) , the age for each of them ( prop . [ fantechiage ] ) and the explicit computation of the orbifold cohomology of @xmath22 , which we recollect in the following two theorems : the orbifold poincar polynomial of @xmath22 is : @xmath23 the compactified orbifold poincar polynomial ( see definition [ pp2 ] ) of @xmath24 is : @xmath25 the ordinary additive cohomology of @xmath22 and @xmath18 was studied by looijenga and getzler , respectively in ( * ? ? ? * ( 4.7 ) ) and ( * ? ? ? 16 ) : our results are an extension of theirs . the paper is developed on three levels of generality . on the first level , we work with a general smooth deligne mumford stack : we recall the theory of inertia stack , age grading , orbifold cohomology , as first developed by chen ruan in @xcite and abramovich graber vistoli in @xcite , @xcite . the compactified orbifold cohomology is introduced for any choice of a smooth compactification @xmath26 . on the second level , we develop the above theories in the case of the moduli stack of curves @xmath9 and its deligne mumford compactification @xmath27 . finally the third level is devoted to work out all the details in the case @xmath1 . we work with cohomology , as in the seminal paper of chen ruan . nevertheless , many of the techniques we use are algebraic , thus closer to the abramovich vistoli s approach . we work over the field of complex numbers , and cohomology is always taken with rational coefficients . [ intro ] by a stack , we shall always mean a deligne mumford stack , of finite type over @xmath28 . in this context , the canonical map from a stack to its coarse moduli space induces an isomorphism in cohomology : we will often identify the two cohomologies by means of this isomorphism . we adopt the convention that orbifold cohomology is the graded vector space underlying chen ruan cohomology , where the latter carries the additional ring structure . if @xmath4 is an algebraic variety , or , more generally , a deligne mumford stack , and @xmath29 is a finite group acting on it , we denote the quotient deligne mumford stack by @xmath30 $ ] . when @xmath4 is a point , we write @xmath31 for @xmath32 $ ] : the moduli stack that classifies principal @xmath29-bundles . in our work , we shall consider the cohomology with its mixed hodge structures . we shall denote by @xmath33 the hodge structure of tate of weight @xmath34 . the class of @xmath35 in the grothendieck group @xmath36 of rational hodge structures will be denoted by @xmath37 $ ] . results on the cohomology with compact support of a quasi - projective variety ( or stack with quasi - projective coarse moduli space ) @xmath4 shall often be expressed by means of its euler characteristic in @xmath36 . following ( * ? ? ? * definition 5.5.2 ) , we call this euler characteristic the _ hodge grothendieck character for compact support _ of @xmath4 and denote it by @xmath38\in k_0({\mathsf{hs}_{\mathbb q}}).\ ] ] hodge grothendieck characters for compact support are sometimes called serre characteristics in the literature . similarly , to state results on cohomology with compact support in a concise way , we shall express them as polynomials with coefficients in the grothendieck group of hodge structures : @xmath39t^i\in k_0({\mathsf{hs}_{\mathbb q}})[t].\ ] ] we work with cyclic covers @xmath40 , where @xmath7 and @xmath41 are , respectively , the covering and the covered space . if a cyclic cover is not tale , the source and the target contain , respectively , ramification and branch points . we shall denote the symmetric group in @xmath42 letters by @xmath43 and the group of @xmath44th roots of unity by @xmath45 . in this section we recollect some basic notions concerning the inertia stack . for a more detailed study , we address the reader to ( * ? ? ? * section 3 ) . we introduce the following natural stack associated to a stack @xmath4 , which points to where @xmath4 fails to be an algebraic space . [ definertia ] ( @xcite , ( * ? ? ? * definition 3.1.1 ) ) let @xmath4 be a stack . the _ inertia stack _ @xmath46 of @xmath4 is defined as : @xmath47 where @xmath48 is the following groupoid : 1 . the objects are pairs @xmath49 , where @xmath50 is an object of @xmath4 over @xmath51 , and @xmath52 is an injective homomorphism ; 2 . the morphisms are the morphisms @xmath53 of the groupoid @xmath54 , satisfying @xmath55 . the inertia stack comes with a natural forgetful map @xmath56 . we also define @xmath57 . the connected components of @xmath58 are called _ twisted sectors _ of the inertia stack of @xmath4 , or simply twisted sectors of @xmath4 . we remark that , by its very definition , @xmath59 is an open and closed substack of @xmath46 , but it rarely happens that it is connected . one special case is when @xmath10 equals @xmath13 : in this case the map @xmath60 restricted to @xmath61 induces an isomorphism of the latter with @xmath4 . the connected component @xmath61 will be referred to as the _ untwisted sector_. we also observe that after the choice of a generator of @xmath62 , we obtain an isomorphism of @xmath46 with @xmath63 , where the latter is defined as the ( @xmath12-)fiber product @xmath64 where both morphisms @xmath65 are the diagonals . [ mappaiota ] there is an involution @xmath66 , which is induced by the map @xmath67 given by @xmath68 . the inertia stack , which we have just defined , is the fundamental ingredient in the definition of orbifold cohomology ( chen ruan cohomology as a vector space ) . we observe that , at this level , we do not need @xmath4 to be smooth nor proper . ( @xcite ) [ defcoomorb1 ] let @xmath4 be a stack . the _ orbifold cohomology _ ( with rational coefficients ) of @xmath4 is defined as a vector space as : @xmath69 now if @xmath70 is an open dense embedding , we can define an intermediate space between @xmath46 and @xmath71 , namely : [ ibar ] given a compactification of a stack @xmath72 , we define the _ compactified inertia stack _ of @xmath4 as the stack @xmath73 where @xmath74 is the stack of all connected components @xmath75 of @xmath76 such that @xmath77 . we can thus define the _ compactified orbifold cohomology _ as the following vector space : @xmath78 in the following sections , we will study the inertia stack for moduli of smooth genus @xmath0 curves and its compactified inertia stack with respect to the deligne mumford compactification . in the first case we will use the theory of cyclic covers of smooth curves ( see @xcite ) , in the second we will use the theory of admissible covers as developed in @xcite . [ stackyremark ] the orbifold cohomology of @xmath4 only depends upon the topological space ( coarse moduli space ) underlying @xmath46 . in @xcite , the authors introduce two notions related to the inertia stack : the _ stack of cyclotomic gerbes _ ( ( * ? ? ? * definition 3.3.6 ) ) and the _ rigidified inertia stack _ ( @xcite ) , showing in @xcite that they are equivalent categories . it is relevant to observe that all these different notions of inertia stacks share the same coarse moduli space , and therefore they give rise to the same orbifold cohomology theory . we want to study the twisted sectors of the inertia stack of moduli of smooth genus @xmath0 curves . for this , we study the moduli stacks of cyclic ramified covers of curves of genus @xmath79 . this approach is due to fantechi @xcite , and builds on the theory of abelian covers of algebraic varieties ( see pardini @xcite ) . the first author has used this approach in ( * ? ? ? * section 2.b ) for the simpler cases when the genus is @xmath12 , or when there is a positive number of marked points . a description of the theory of abelian covers in the case of curves and of cyclic groups that is closely related to the one we use can be found in ( * ? ? ? * sections 1 and 2 ) . a similar construction for covers of prime order was studied in @xcite . we start by summarizing informally the description of cyclic covers we will use in this paper . [ pardiniprop ] ( ( * ? ? ? * proposition 2.1 ) ) let @xmath41 be a smooth genus @xmath80 curve . then the following data are equivalent : * a cyclic ( possibly ramified ) @xmath62-cover @xmath81 , where @xmath7 is a smooth curve , possibly disconnected ; * a sequence of @xmath11 smooth effective divisors @xmath82 ( with pairwise disjoint support , possibly empty ) , a line bundle @xmath83 on @xmath41 together with an isomorphism @xmath84 . with this result in mind , let us define : [ admissible ] let @xmath85 be an integer . a @xmath0-admissible datum is an @xmath86-tuple of nonnegative integers @xmath87 with @xmath88 and @xmath89 , satisfying the following conditions : * riemann hurwitz formula @xmath90 * the structural equation of abelian covers @xmath91 the integers @xmath10 and @xmath80 will be called respectively the _ order _ and the _ base genus _ of the @xmath0-admissible datum @xmath92 . note that , for a fixed @xmath0 , the set of all @xmath0-admissible data @xmath92 is finite . with every admissible @xmath0-datum , we associate the integers @xmath93 , and a disjoint union decomposition @xmath94 by : @xmath95 moreover , we will denote by @xmath96 the subgroup of @xmath43 ( the symmetric group on @xmath42 elements ) defined by @xmath97 . we shall now construct the twisted sectors of @xmath9 as stacks of cyclic @xmath10-covers of curves of genus @xmath80 , with branch locus of type @xmath98 . [ settoretwistato ] let @xmath92 be a @xmath0-admissible datum . we define the stack @xmath99 whose objects over a scheme @xmath51 are @xmath100-tuples @xmath101 , where @xmath7 is a smooth family of genus @xmath80 curves over @xmath51 , the @xmath102 are sections of @xmath103 ( where @xmath104 denotes the big diagonal ) defining disjoint divisors on @xmath7 , @xmath83 is a line bundle and @xmath105 is an isomorphism . the stack @xmath106 is defined as the open and closed substack of @xmath99 whose objects under the correspondence described in fact [ pardiniprop ] are connected covers . [ condizioneconnessione ] let @xmath92 be a @xmath0-admissible datum , and define @xmath44 as the greatest common divisor of @xmath10 and all the @xmath107 with @xmath108 . the condition of connectedness for the cover in the last sentence of definition [ settoretwistato ] is satisfied when @xmath109 has precisely order @xmath44 in the picard group of @xmath7 . when @xmath80 equals @xmath110 , then @xmath99 is always connected , thus @xmath106 is empty when @xmath111 , or it coincides with @xmath99 when @xmath112 . when @xmath113 , fixing a proper divisor of @xmath44 as the order of determines an open and closed substack of @xmath99 . as a consequence of theorem [ connessione ] , we see _ a posteriori _ that each of these open and closed substacks ( in particular also @xmath106 ) is in fact connected . [ esisteolofisso ] let us be more explicit about the morphisms of @xmath114 . let @xmath115 @xmath116 and @xmath117 be two objects as in definition [ settoretwistato ] . then a morphism between them is a couple of isomorphisms @xmath118 satisfying the following conditions : the map @xmath119 is an isomorphism of curves such that @xmath120 and @xmath121 is an isomorphism of line bundles that makes the following diagram commute : @xmath122^{\hspace{-0.5cm}\sigma^(\phi ' ) } \ar[d]^{\tau^{\otimes n } } & \sigma^\left ( \mathcal{o}_{c ' } ( \sum i d_i')\right ) \ar[d]^{\gamma } \\ l^{\otimes n } \ar[r]^{\hspace{-0.5cm}\phi } & \mathcal{o}_c(\sum i d_i),}\ ] ] where we denoted by @xmath123 the isomorphism induced by @xmath119 . a different definition for a morphism between the two families @xmath124 and @xmath117 is the following : a single isomorphism @xmath125 satisfying @xmath120 and such that there exists an isomorphism @xmath126 making diagram commute . these two different definitions of morphisms give rise to two different stacks , which share the same coarse moduli space . let us denote by @xmath127 the open substack of the moduli stack of stable maps @xmath128 whose source curve is smooth . the moduli stack @xmath128 is defined in @xcite ; see also @xcite , where @xmath128 is denoted @xmath129 . we observe that the stack @xmath99 we have just defined ( with the first definition of morphisms in remark [ esisteolofisso ] ) is an open and closed substack of the quotient stack @xmath130 $ ] prescribed by the assignment of the ramifications @xmath98 . we now show that the moduli stacks @xmath131 we have just constructed constitute open and closed substacks of the inertia stack of @xmath9 . [ corrispondenza ] let us fix @xmath132 . then the stack @xmath133 of definition [ definertia ] is isomorphic to the disjoint union of all nonempty stacks @xmath106 for all @xmath0-admissible data @xmath134 with order @xmath135 equal to @xmath10 . follows from definitions [ admissible ] , [ settoretwistato ] and by adapting the proof of ( * ? ? ? * theorem 2.1 , proposition 2.1 ) to this relative case ( cf . fact [ pardiniprop ] ) . indeed , there is a base - preserving equivalence of categories : @xmath136 where in the right hand side the disjoint union is taken over all @xmath0-admissible data with order @xmath10 . we sketch the proof of this well known fact , by explicitly defining the ( functorial ) correspondence . let us assume that a @xmath62-cover @xmath137 is given ( over a base @xmath51 ) . over each point @xmath138 , the branch divisor @xmath139 of the cover @xmath140 can be split according to local monodromy in smooth effective divisors @xmath141 , having pairwise disjoint support . identifying these divisors with sections of the appropriate symmetric product of @xmath142 gives the @xmath102 . each @xmath102 defines a codimension @xmath13 subscheme of @xmath7 which does not contain any fibre of @xmath142 , hence they give rise to effective cartier divisors on @xmath7 . at this point , we only need to construct the line bundle @xmath83 together with the isomorphism @xmath143 . since the cover is nontrivial , the action of @xmath62 on the push - forward sheaf @xmath144 defines a splitting as a direct sum of line bundles : @xmath145 where @xmath146 is the subsheaf of @xmath144 of sections where @xmath62 acts with weight @xmath107 . the line bundle @xmath83 is then defined as @xmath147 . by viewing the sections of @xmath148 as functions on the total space of the line bundle @xmath149 , and hence also as functions on its trivial section @xmath7 , we obtain an identification @xmath150 the correspondence just defined is essentially surjective . indeed , if the line bundle @xmath83 is given over @xmath41 , we set @xmath151 . the line bundles @xmath152 can then be defined as : @xmath153 } \quad a=0,\ldots , n-1\ ] ] now the number @xmath154 + [ \frac{b i } { n}]- [ \frac{(a+b ) i } { n}]$ ] can either be @xmath110 or @xmath13 . in both cases , the canonical sections of the line bundles : @xmath155 + [ \frac{b i } { n}]- [ \frac{(a+b ) i } { n}]}\ ] ] permit the definition of a ring structure over @xmath156 . now the normalization of the spectrum of @xmath157 reconstructs the smooth @xmath62-cover of @xmath41 up to isomorphism of @xmath62-cover . if one considers the definition of morphisms in the groupoid @xmath158 given in remark [ esisteolofisso ] , one can also check that the correspondence is fully faithful , hence an equivalence of categories . the moduli spaces @xmath106 are indeed connected . this would follow , after some work , from ( * ? ? ? * theorem 1.2 ) . we refer to a more recent work of catanese , where the result we need is stated in the same language used in this paper . ( ( * ? ? ? * theorem 2.4 ) ) [ connessione ] let @xmath92 be a @xmath0-admissible datum . when the stack @xmath106 is nonempty , it is connected . in particular , as a consequence of this connectedness result , the nonempty moduli stacks @xmath106 give all the twisted sectors of the inertia stack of @xmath9 . working with the second definition in remark [ esisteolofisso ] of morphism for the stack @xmath99 , one obtains a decomposition of the rigidified inertia stack ( see remark [ stackyremark ] ) of @xmath9 . in this section , we study the geometry of the twisted sectors of the inertia stack of the moduli space of smooth , genus @xmath2 curves . we determine the cohomology of all these twisted sectors as a graded vector space with hodge structures . we shall state these results in the form of polynomials with coefficients in @xmath36 ( see section [ notation ] ) . our approach is based on the correspondence between twisted sectors and @xmath0-admissible data introduced in the previous section . of course , in the case of genus @xmath2 also a direct approach is possible by classifying all automorphisms of plane quartic curves ( as in e.g. ( * ? ? ? * lemma 6.5.1 ) ) and of all hyperelliptic genus @xmath2 curves . however , our approach seems more suitable for cohomological computations and has the advantage that it generalizes to higher genus . if @xmath4 is a twisted sector of @xmath159 , we have seen in the previous section that @xmath160 for @xmath92 a certain @xmath2-admissible datum . we start by considering the admissible data with @xmath161 . [ basegenus0 ] there are @xmath162 different @xmath2-admissible data @xmath92 with @xmath161 that parametrize connected covers . the complete list of these admissible data and of the hodge grothendieck characters for compact support of the associated twisted sectors @xmath163 is given in tables [ base0dim>0 ] and [ base0dim0 ] . .[base0dim>0]positive - dimensional twisted sectors . for the sake of brevity we omit @xmath161 from the notation of the admissible datum . [ cols="^,^,^,^ " , ] now we describe one by one the irreducible components of @xmath164 that map onto the four codimension @xmath13 boundary strata we have just pictured : 1 . there are two irreducible components @xmath165 lying over @xmath166 . they parametrize admissible double covers @xmath167 such that the restriction to the preimage of the genus @xmath13 component is , respectively , a trivial @xmath168-cover in the case of @xmath169 and a nontrivial @xmath168-cover in the case of @xmath170 . the coarse moduli space of @xmath169 is isomorphic to @xmath171 , and its hodge grothendieck character for compact support is then @xmath172 . the moduli space of @xmath170 is isomorphic to @xmath173 and @xmath174 . 2 . over @xmath175 there is one irreducible component @xmath176 , whose moduli space is isomorphic to @xmath177 . here @xmath178 is the moduli space of bielliptic curves with a choice of a distinguished bielliptic involution , and an ordering of the ramification points ( see ( * ? ? ? * proposition 2.25 ) ) . as shown there , the stack @xmath178 has @xmath179 as coarse moduli space . hence , one has @xmath180 ; 3 . over @xmath181 there is one component @xmath182 . its moduli space is @xmath183 , where @xmath184 is the moduli space of bielliptic curves @xmath7 with a distinguished bielliptic involution @xmath8 and a point @xmath185 not fixed by @xmath8 , and the involution on @xmath184 is defined by sending @xmath186 to @xmath187 . using the construction of ( * ? ? ? * proposition 2.22 ) , it is easy to show that the coarse moduli space of @xmath183 is a @xmath188-bundle over @xmath179 . in particular , one has @xmath189 ; 4 . finally , over @xmath190 there are two components , one whose general element is a cover unramified over the node , and the other one whose general element is totally ramified over the node . both these moduli spaces are isomorphic to @xmath191 , and their hodge grothendieck character for compact support is then @xmath192 . the second step is to study the number of irreducible components of the preimages in @xmath193 of each of the codimension @xmath12 strata of @xmath194 $ ] . there are exactly nine codimension @xmath12 strata @xmath195 in @xmath196 $ ] . we describe them in figure [ cod2mb14 ] by drawing their general element . ccc the nine strata of codimension @xmath12 in @xmath197$],title="fig : " ] & the nine strata of codimension @xmath12 in @xmath197$],title="fig : " ] & the nine strata of codimension @xmath12 in @xmath197$],title="fig : " ] + @xmath198&@xmath199&@xmath200 + + the nine strata of codimension @xmath12 in @xmath197$],title="fig : " ] & the nine strata of codimension @xmath12 in @xmath197$],title="fig : " ] & the nine strata of codimension @xmath12 in @xmath197$],title="fig : " ] + @xmath201&@xmath202&@xmath203 + the nine strata of codimension @xmath12 in @xmath197$],title="fig : " ] & the nine strata of codimension @xmath12 in @xmath197$],title="fig : " ] & the nine strata of codimension @xmath12 in @xmath197$],title="fig : " ] + @xmath204&@xmath205&@xmath206 * the strata @xmath207 and @xmath200 lie in the closure of the codimension @xmath13 stratum @xmath166 . again , the preimage of each of them under @xmath208 has two components , corresponding respectively to trivial and nontrivial @xmath168-covers of the irreducible component of genus @xmath13 ; * the stratum @xmath201 is contained in the closure of @xmath175 and its preimage under @xmath208 is irreducible ; * the preimage @xmath209 has two components , corresponding to whether the two nodes are contained or not in the branch locus ; * the preimage @xmath210 is irreducible ; * the strata @xmath211 and @xmath206 lie in the closure of @xmath190 and the preimage of each of them has @xmath12 irreducible components . indeed , while the fact that the separating node is or not part of the branch locus depends on the disposition of the other branch points , the irreducible node can be either a branch point or a regular point , which gives rise to two different components , exactly as it was the case for @xmath190 so , putting everything together , we have : @xmath212 and now , using the fact that poincar duality holds , we get to the conclusion : @xmath213 in the proof of the above result we have used poincar duality . it is possible to reach the same conclusion by means of a more refined analysis of the boundary strata . this gives a non - trivial check on the whole result . more precisely , the quantity @xmath214 is obtained as the sum of the following contributions : @xmath215 where we have collected the terms according to the codimension of the corresponding stratum ( from codimension @xmath110 to codimension @xmath216 ) . [ pscb ] the hodge grothendieck character of @xmath217 is @xmath218 . we have seen in corollary [ psb ] that @xmath219 equals @xmath220 . the moduli stack @xmath221 admits a finite tale map onto @xmath222 . on its compactification @xmath217 by means of admissible covers this map extends to a finite map @xmath223 . the boundary @xmath224 has two irreducible components : namely , a component @xmath225 whose general element is of compact type , and another component @xmath226 whose general element is an irreducible curve of geometric genus @xmath110 with one node . it is easy to see that the preimage of both @xmath226 and @xmath225 under @xmath208 has two irreducible components . by the additivity of hodge grothendieck characters for compact support this ensures that the coefficients of @xmath83 and @xmath227 in the hodge grothendieck character of @xmath228 are as in the claim . the constant coefficient is equal to the degree @xmath12 coefficient by poincar duality . [ pscc ] the hodge grothendieck character of @xmath229 is @xmath218 . we have seen in proposition [ psc ] that @xmath230 equals @xmath220 . the moduli stack @xmath231 admits a finite tale map onto @xmath232 $ ] . this map extends to a finite map @xmath233 $ ] on the compactification @xmath234 by means of admissible covers . there are two irreducible components of codimension @xmath13 in @xmath235 $ ] , one whose general element is of compact type , and the other one whose general element is an irreducible curve of geometric genus @xmath110 with one node . we call them respectively @xmath225 and @xmath226 . now the fiber of @xmath208 over @xmath225 is made of one irreducible component , while the fiber of @xmath208 over @xmath226 is made of three irreducible components , one for each possible way of prescribing a balanced ramification over the node . [ pscd ] the stack @xmath236 has hodge grothendieck character for compact support @xmath237 one can compute the hodge grothendieck character of @xmath236 with the same strategy used for the propositions [ psca][pscc ] , but also a more geometric proof is possible : as shown by bernstein in ( * ? ? ? * sections 2.52.6 ) , the coarse moduli space of admissible tale double covers in genus @xmath0 coincides with the coarse _ moduli space of prym curves _ , i.e. of quasistable curves together with a line bundle which is isomorphic to @xmath238 on the exceptional components and to a square root of the trivial bundle on the nonexceptional ones . indeed the latter coarse moduli space , usually denoted by @xmath239 , has been extensively studied . the results of bernstein imply that @xmath236 has coarse moduli space @xmath240 , and by ( * ? ? ? * lemma 20 ) the latter coincides with @xmath241 . the combination of these results implies proposition [ pscd ] . to complete our computation of the orbifold cohomology of @xmath3 we still need to consider its structure as a @xmath242-graded vector space . to compute the new grading , the degree of each cohomology class of the inertia stack of @xmath3 has to be shifted by the _ age _ of the twisted sector . we recall from the introduction that the motivation for the new grading is that orbifold cohomology can be endowed with a natural product , the chen ruan product , which is not compatible with the `` naive '' grading of the cohomology of the inertia stack , but turns it into a @xmath242-graded algebra , once it is endowed with the new grading . ( ( * ? ? ? * section 7.1 ) ) let @xmath245 be a group homomorphism . it is determined by an integer @xmath246 as @xmath247 . we define the _ age _ of @xmath248 by : @xmath249 the age extends to a unique additive homomorphism @xmath250 . [ definitionage ] ( ( * ? ? ? * section 3.2 ) , ( * ? ? ? * definition 7.1.1 ) ) let @xmath75 be a twisted sector and @xmath252 a point . then the pull - back via @xmath253 of the tangent sheaf , @xmath254 , is a representation of @xmath62 on a finite - dimensional vector space . we define : @xmath255 [ pp1 ] we define the _ orbifold poincar polynomial _ of @xmath4 as : @xmath257 \cap \mathbb{q } } \dim h^{i}_{cr}(x ) t^i\ ] ] note that the degree of @xmath258 is given by the unconventional grading defined in definition [ defcoomorb2 ] . in definition [ ibar ] we have introduced a compactification of the inertia stack . the connected components of @xmath259 can be assigned the age grading as in this section , simply using the fact that every connected component of @xmath259 contains , by its very definition , a connected component of @xmath260 . taking cohomology , one obtains a linear subspace of the orbifold cohomology @xmath261 . although we shall not deal with this in the present paper , it is actually possible to prove that the orbifold cohomology classes coming from @xmath259 form a subalgebra of the chen ruan cohomology ring . [ pp2 ] we define the compactified orbifold cohomology @xmath262 as @xmath263 with the grading induced from @xmath264 . the additive structure of this vector space can be recollected in the polynomial : @xmath265 \cap \mathbb{q } } \dim \overline{h}^{i}_{cr}(x ) t^i\ ] ] we call this polynomial the _ compactified orbifold poincar polynomial_. we observe that poincar duality holds for the orbifold cohomology of @xmath4 if @xmath4 is a proper smooth stack , and for the compactified orbifold cohomology of @xmath4 with respect to the compactification @xmath266 , if @xmath267 is smooth . [ fantechiage ] ( @xcite ) let @xmath85 and let @xmath75 be the twisted sector of @xmath268 corresponding to the discrete datum @xmath269 ( definition [ admissible ] , proposition [ corrispondenza ] ) . then its age is equal to : @xmath270 where @xmath271 if @xmath272 and @xmath13 otherwise . a point of @xmath273 is a @xmath62-cover @xmath274 . by standard infinitesimal deformation theory , the fiber of the tangent sheaf @xmath275 at @xmath7 is @xmath276 . thus the cyclic group @xmath62 acts on the vector space @xmath276 , and the latter splits in a direct sum of eigenspaces as @xmath277 where @xmath278 denotes the subspace where @xmath62 acts with weight @xmath44 . according to definition [ definitionage ] , we have : @xmath279 now by stability we can substitute @xmath280 with @xmath281 . moreover , since @xmath282 is finite and @xmath283 is coherent , we have that all higher invariant direct images @xmath284 vanish ( see ( * ? ? ? * chapter v , corollary p. 202 ) ) , and thus : @xmath285 so becomes @xmath286 the sheaf @xmath287 is studied in ( * ? ? ? * proposition 4.1,(a ) ) . in particular , one has @xmath288 where the degree of the line bundles @xmath289 can be computed via ( * ? ? ? * proposition 2.1 ) ( see also ): @xmath290 note that by combining proposition [ fantechiage ] with ( * ? ? ? * lemma 4.6 ) ( respectively , with ( * ? ? ? * corollary 4.10 ) ) , one obtains a closed formula for the age of the twisted sectors of @xmath14 ( respectively , @xmath292 , the moduli space of stable curves that have one smooth component of geometric genus @xmath0 ) . the ordinary poincar polynomial of @xmath22 is calculated in ( * ? ? ? * ( 4.7 ) ) . then the result follows from the determination of the poincar polynomials of the moduli stacks @xmath106 carried out in section [ inertiam3 ] for any @xmath2-admissible datum @xmath92 . see in particular table [ base0dim>0 ] and [ base0dim0 ] , proposition [ psa ] and [ psc ] and corollary [ psb ] and [ psd ] . in our work we have computed the hodge structures on the twisted sectors of @xmath3 , which turned out to be always pure and of tate type . therefore , if following ( * ? ? ? 3.2.4 ) one defines @xmath294 where the sum runs over all sectors @xmath75 of @xmath3 , one can endow @xmath295 with a canonical structure of direct sum of pure hodge structures of tate type whose hodge filtrations have been shifted by a rational number . under this convention , what we actually prove is the following result : @xmath296 the ordinary poincar polynomial of @xmath18 was first computed in ( * ? ? ? the proof of this theorem follows again as a recollection of the results obtained in section [ compm3 ] , in particular the propositions [ psca ] , [ pscb ] , [ pscc ] , [ pscd ] for the poincar polynomials of @xmath164 , @xmath217 , @xmath229 , @xmath297 respectively . for the remaining twisted sectors , the poincar polynomial is computed applying corollary [ corollariocadman ] and the results are summarized in table [ base0dim>0 ] and [ base0dim0 ] . dan abramovich , tom graber , angelo vistoli , _ algebraic orbifold quantum products _ , orbifolds in mathematics and physics ( madison , wi , 2001 ) , 124 , contemp . , 310 , amer . soc . , providence , ri , 2002 . weimin chen , yongbin ruan , _ a new cohomology theory of orbifold _ ( 2004 ) , no . maurizio cornalba , _ on the locus of curves with automorphisms _ , ann . pura appl . ( 4 ) 149 ( 1987 ) , 135151 . eduard looijenga , _ cohomology of @xmath22 and @xmath303 _ , mapping class groups and moduli spaces of riemann surfaces ( gttingen , 1991/seattle , wa , 1991 ) , 205228 , contemp . , 150 , amer . soc . , providence , ri , 1993 .	in this work we study the additive orbifold cohomology of the moduli stack of smooth genus @xmath0 curves . we show that this problem reduces to investigating the rational cohomology of moduli spaces of cyclic covers of curves where the genus of the covering curve is @xmath0 . then we work out the case of genus @xmath1 . furthermore , we determine the part of the orbifold cohomology of the deligne mumford compactification of the moduli space of genus @xmath2 curves that comes from the zariski closure of the inertia stack of @xmath3 .
the evolution of clusters of galaxies has important implications to our knowledge of the fundamental cosmological parameters . as they are the most massive gravitationally bound systems in the universe , the evolution of clusters is highly sensitive to the physics of cosmic structure formation and the values of the fundamental cosmological parameters . moreover , clusters are valuable laboratories to study the evolution of many galaxies at a common redshift . although a number of galaxies with luminosities @xmath15 are now observed out to redshifts @xmath16 ( @xcite ) , very few candidate clusters of galaxies have been spectroscopically confirmed at redshifts near or above unity ( @xcite ) . at these very high redshifts , the prevalence of clusters , proto - clusters , or other large scale structures in the galaxy distribution is as yet unknown , and the evolution of large scale structures may well be at a critical stage ( @xcite ) , where observations can directly constrain cosmological models . at intermediate redshifts , powerful radio - galaxies are frequently located in rich clusters ( @xcite ) , therefore , a possible search strategy is to look for clusters around known powerful radio - galaxies . as part of a program to identify clusters and study their properties at redshifts larger than @xmath17 , we report here the discovery of galaxies within @xmath18 mpc and 1000 kms@xmath13 from the 1jy radio galaxy mrc0316 - 257 ( @xcite ) , at @xmath19 . @xmath20 kms@xmath13mpc@xmath13 and @xmath21 are used throughout this letter . we have observed a @xmath22 field around mrc0316 - 257 with the multi object imaging spectrograph ( mos ) at cfht ( @xcite ) . deep broad band v and i images as well as narrow band images in a filter with central wavelength 5007 and bandwidth 96 , containing the @xmath7 line redshifted to @xmath0 , were obtained the night of 20 december 1995 . several images were obtained for each filter , after small telescope offsets were performed , to improve the flat fielding accuracy . total integration times were 2700 , 1200 , 5100 seconds in v , i and the 5007 filters , respectively . standard image processing was performed the next day , and the v and 5007 images were blinked visually to provide a list of cluster candidates for subsequent multi - slit spectroscopy during the nights 21 - 24 december 1995 . the narrow band 5007 images can be expected to identify galaxies with @xmath7 in a redshift range @xmath23 , foreground galaxies with [ oii]3727 in emission with a redshift @xmath24 , or the rarer occurence of galaxies with [ oiii]5007 with @xmath25 . the narrow band image is compared to the v band image of the mrc0316 - 257 field in figure 1 ( plate 1 ) . two galaxies exhibit strong excess emission in the narrow band filter , and were subsequently observed spectroscopically , together with more marginal candidates . multi - slit masks were prepared with the highest priority given to the cluster candidates as defined above , and other slits were placed on galaxies brighter than @xmath26 when no other candidate was available . the observations of about 30 galaxies per mask were performed with the o300 grism , with 300 l / mm and peak transmission at @xmath275900 , and slits @xmath28 wide , providing a spectral range of 4500 - 9000 and a resolution of @xmath29 . three masks have received cummulated exposure times of 19800 , 8700 , and 8700 seconds . data reduction of the spectra in the first mask obtained the night of 21 december 1995 was performed the following afternoon , and confirmed that these candidate galaxies had emission lines at @xmath275030 . these galaxies were subsequently reobserved in the following multi - slit mask for a total exposure time of 28500 sec . final data reduction was performed with the multired package implemented under iraf , and redshifts were measured and assigned classes as described in le fvre et al . the two galaxies common to 2 different masks were treated separately by averaging the flat - fielded and sky corrected 2d spectra obtained from each 2 mask . the @xmath30 detection limit in the shortest integration spectra is @xmath31 erg@xmath12sec@xmath13@xmath13 around 5000 , and our spectra would have allowed us to identify any emission line with a flux larger than @xmath32 erg@xmath12sec@xmath13 . a total of 51 galaxies have secure redshift measurements in the range @xmath33 ; 2 stars were identified , and we failed to measure redshifts for 45 galaxies . mccarthy et al . ( 1990 ) noted that @xmath7 was the only feature in support of their redshift for mrc0316 - 257 , while eales et al . ( 1993 ) identified [ oiii]5007 from ir spectroscopy . our spectra confirm @xmath7 in emission , as well as civ1549 , heii1640 and ciii]1909 ( figure 1 ) . the redshift of the radio - galaxy mrc0316 - 257 is therefore @xmath2 . the v image shows another galaxy of roughly equal magnitude 2.7 arcseconds away , while in the 5007 image the eastern galaxy disappears almost completely . in addition to the radio - galaxy , we have identified spectroscopically two galaxies with a redshift similar to the radio - galaxy . the first one , `` galaxy a '' , is 44 arcseconds away from the radio - galaxy . its spectrum shows a strong line at 5030.3 , and faint features with marginal s / n at the same redshift ( fig.2 ) . a cross correlation of the spectrum of galaxy a , after subtraction of the 5030 emission line , with a template spectrum with line widths taken from @xcite ( 1996 ) , gives a marginal correlation at z=3.14 with a correlation coefficient r=2.8 . the observed equivalent width of the emission line is 230@xmath34 , which , if it was [ oii]3727 at z=0.350 , would indicate a rest frame @xmath35 , larger than any equivalent width for galaxies in the canada - france redshift survey ( cfrs ) , or 3 times larger than for any galaxies at @xmath36 in the cfrs ( @xcite ) . moreover , all galaxies in the cfrs with rest @xmath37 have rest @xmath38 and @xmath39 , which would have been easily identified in our spectra at a redshift @xmath40 . the identification of the line at 5030 with @xmath7 at @xmath3 seems therefore to be secure . the second galaxy , `` galaxy b '' , is located 182 arcseconds from mrc0316 - 257 . its spectrum shows @xmath7 and civ1549 , and therefore gives a secure redshift @xmath4 . both civ1549 , and to a lesser extent @xmath7 exhibit a broad component , which indicates the presence of an agn . close examination of the images indicate that both mrc0316 - 257 and `` galaxy a '' are resolved under fwhm=1.5 arcseconds seeing , while `` galaxy b '' is only marginally resolved . after deconvolution from the observed psf , the half - light radii of mrc0316 - 257 and `` galaxy a '' , are @xmath41 and @xmath42 @xmath43 kpc respectively . the fact that galaxy b is quite compact supports the contribution of an agn . the v , i and @xmath7 images of galaxy a are strinkingly different : the peaks of continuum v and line emission are offset by 0.6 arcsec , with the peak of @xmath7 emission extending toward the sw , where little or no continuum light is detected in the v and i bands , and little or no emission is observed in i at the peak of @xmath7 emission , while several blobs of emission appear 3 arcseconds to the nw in the i band . in figure 3 we plot the ( v@xmath44i ) vs ( v@xmath445007 ) color - color diagram . besides mrc0316 - 257 , and the two associated galaxies identified in spectroscopy , two galaxies have significant excess emission in the 5007 filter ( v@[email protected] ) . one of these galaxies has a redshift @xmath46 , and the excess emission in the 5007 filter is coming from the [ oii]3727 emission line . the other galaxy , galaxy c , has not been observed spectroscopically . therefore , in our imaging field of view of 78.2 arcmin@xmath47 , we have three confirmed galaxies at @xmath0 , above our detection thresholds of @xmath48 and @xmath49 , and a fourth galaxy , galaxy c , which remains a good but unconfirmed candidate galaxy at @xmath0 . any estimate of the space density of @xmath50 galaxies from our observations is necessarily quite uncertain due to the small number of objects which we detect . the maximum redshift depth probed is set by the range over which our narrow band filter would detect the ly@xmath1 emission line , i.e. @xmath51 . under that assumption , the resulting co - moving cosmological volume surveyed is therefore @xmath52mpc@xmath53 assuming @xmath54 . ( for @xmath55 , the effective volumes are 7.4@xmath56 larger , and the densities correspondingly smaller . ) however , the three galaxies which we have confirmed spectroscopically are confined to a much smaller redshift range , close to the radio galaxy : @xmath57 . this corresponds to a rest frame velocity difference @xmath58 km s@xmath13 , quite in line with expectations for galaxy clusters and/or `` walls '' of large scale struture . it is therefore likely that the effective `` cluster '' volume is much smaller than the value given above . assuming @xmath59 , the corresponding co - moving volume would then be @xmath60mpc@xmath53 . we therefore may bracket the range of possible space densities by assuming @xmath61 ( mpc@xmath62 , or @xmath63mpc@xmath64 , with a `` best guess '' value of @xmath65mpc@xmath66 . the population of @xmath67 field galaxies identified by steidel _ _ 1996 has a comoving space density of @xmath68mpc@xmath66 for @xmath69=0.5 . at face value , this would imply that our `` best guess '' space density represents an overdensity of @xmath70 compared to the field population at similar redshift . however , it is difficult to make a direct comparison between our `` cluster '' space density and the``field '' value since our galaxies were selected on the basis of strong ly@xmath1 flux , and have brighter continuum magnitudes ( @xmath49 ) than the steidel _ _ objects ( which have @xmath71 @xmath72 ) . therefore the estimate given above for the cluster overdensity is almost certainly a substantial _ underestimate . _ even our derived lower limit of @xmath73mpc@xmath66 for the 0316 - 257 field is substantially higher than the corresponding space density of field galaxies sharing the same continuum magnitudes and ly@xmath1 properties . the rest frame @xmath7 equivalent widths are @xmath74 , @xmath75 , @xmath76 , for the radio galaxy , and galaxies a and b respectively . while the radio - galaxy and galaxy b have resolved @xmath7 , with @xmath77 and @xmath78 respectively , and civ is resolved for galaxy b , with @xmath79 after deconvolution by the instrumental response , @xmath7 is unresolved at our resolution for galaxy a. this may indicate that @xmath7 in galaxy a is produced mainly by stellar photo - ionisation . even after correcting for the contribution of the lyman alpha line to the v band , galaxy a is extremely blue , with @xmath80 . this is very unusual for faint galaxies , and indeed galaxy a is by far the bluest object in our field of view , as can be seen from figure 3 . galaxy a is bluer than most or all of the z@xmath273 objects reported by steidel et al . it is also interesting that its lyman alpha emission is significantly stronger than that in most of the steidel et al . lyman break galaxies . the blue color and strong ly - a emission may indicate an object dominated by ob - stars in a strong starburst with very little dust . the complex morphology of galaxy a is quite different from the morphology of galaxies identified on the basis of the 912 discontinuity ( @xcite , @xcite ) , with complex spatial distribution of continuum and line emission as described in section 3 . while the evidence for old stars will have to be searched for at redder wavelength , this galaxy might well be forming its first stars and fit the definition of a proto - galaxy . we have spectroscopically identified 2 galaxies at the same redshift as the radio - galaxy mrc0316 - 257 , z@xmath81 . these galaxies exhibit @xmath7 in emission with @xmath82 . one galaxy is resolved with a size @xmath83 mpc , has unresolved @xmath7 , and extremely blue @xmath9 color , which might indicate that the main photo - ionising process is coming from a first generation of young stars in a low dust medium . the other galaxy is unresolved , and has broad @xmath7 and civ emission lines , indicating the presence of an agn . these observations of galaxies around mrc0316 - 257 provide tentative evidence for a region of high galaxy density , possibly indicative of clustering of galaxies occuring at @xmath19 . they indicate that the search for galaxies around very high redshift radio - galaxies is indeed possible from @xmath7 imaging techniques . however , the total number of galaxies identified with an observed @xmath7 luminosity @xmath11 erg@xmath12sec@xmath13 is small , indicating the limits of this technique , in agreement with previous searches . lyman break techniques , as successfuly applied by steidel et al . ( 1996 ) , will be required to identify additional galaxies at the same redshift and confirm the existence of a cluster or proto - cluster of galaxies . bruzual , a.g . , 1983 , apjs , 53 , 497 dickinson , m. , 1994 , phd thesis dickinson , m. , 1995 , in `` galaxies in the young universe '' , eds . h. hippelein , k. meisenheimer , & h - j . rser , springer , p.144 . eales , s.a . , & rawlings , s. , 1993 , , 411 , 67 evrard , a.e . , & charlot , s. , 1994 , , 424 , l14 fontana , a. , cristiani , s. , dodorico , s. , giallongo , e. , savaglio , s. , 1996 , , 279 , l27 francis , p.j . , woodgate , b.e . , warren , s.j . , moller , p. , mazzolini , m. , bunker , a.j . , lowenthal , j.d . , williams , t.b . , minezaki , t. , kobayashi , y. , yoshii , y. , 1996 , , 457 , 490 frenk , c.s . , evrard , a.e . , white , s.d.m . , summers , f.j . , 1996 , , preprint giavalisco , m. , steidel , c.c . , & szalay , a.s . , 1994 , 425 , l5 giavalisco , m. , steidel , c.c . , & macchetto , f. , 1996 , , in press hammer , f. , flores , h. , lilly , s.j . , crampton , d. , le fvre , o. , rola , c. , mallen - ornelas , g. , schade , d. , tresse , l. , , preprint hill , g.j . , lilly , s.j . , 1991 , , 367 , 1 le fvre , o. , crampton , d. , hammer , f. , lilly , s.j . , tresse , l. , 1994a , , 424 , l14 le fvre , o. , crampton , d. , felenbok , p. , monnet , g. , 1994b , , 282 , 325 le fvre , o. , crampton , d. , lilly , s.j . , hammer , f. , tresse , l. , 1995 , , 455 , 60 mccarthy , p.j . , kapahi , v.k . , van breugel , w. , subrahmanya , c.r . , 1990 , , 100 , 1014 pascarelle , s.m . , windhorst , r.a . , driver , s.p . , ostander , e.j . , & keel , w.c . , 1996 , , 456 , l21 peebles , p.j.e . , daly , r. , juszkiewicz , 1989 , , 347 , 563 steidel , c.c . , giavalisco , m. , pettini , m. , dickinson , m. , adelberger , k. , 1996 , , in press yee , h. , green , 1984 , , 280 , 79 yee , h. , ellingson , e. , bechtold , j. , carlberg , r. , cuillandre , j.c . , 1996 , aj , in press ( may 1996 ) lccccccc 0316 - 257 & @xmath84 @xmath85 & @xmath86 @xmath85 & [email protected] 10@xmath88 & [email protected] & [email protected] & @xmath4423.27 & 8.9@xmath89 a & @xmath90 & @xmath91 & [email protected] 10@xmath88 & [email protected] & [email protected] & @xmath4422.67 & 11.6@xmath92 b & @xmath93 & @xmath94 & [email protected] 10@xmath88 & [email protected] & [email protected] & @xmath4423.30 & 2.3 @xmath95 c @xmath96 & @xmath97 & @xmath98 & 0.5 10@xmath88 @xmath99 & [email protected] & [email protected] & @xmath4422.64 & 6.3@xmath100	we report here the spectroscopic identification of galaxies in the neighborhood of the radio - galaxy mrc0316 - 257 , at a redshift @xmath0 . candidate cluster galaxies were selected from deep v and i images combined with narrow band imaging at the wavelength of redshifted ly@xmath1 . follow - up multi - slit spectroscopy has allowed confirmation of the redshift of the radio - galaxy , @xmath2 , and identification of two associated galaxies at redshifts @xmath3 and @xmath4 respectively . the first galaxy is 0.3 @xmath5 mpc from the radio - galaxy , is resolved with an intrinsic size @xmath6 kpc , and shows @xmath7 in emission with rest @xmath8 . in addition , its extremely blue @xmath9 color might possibly indicate a proto - galaxy forming a first generation of stars in a low dust medium . the second galaxy is 1.3 @xmath5 mpc away from the radio - galaxy , is marginally resolved and , in addition to ly@xmath1 in emission , shows civ in emission with a broad component indicating the contribution from an agn . the comoving density of galaxies with @xmath10 and a @xmath7 flux @xmath11 erg@xmath12sec@xmath13 in the vicinity of mrc0316 - 257 is @xmath14 , significantly higher than the expected background density of field galaxies with similar properties , and might indicate a rich cluster or proto - cluster environment . these observations indicate that the environment of high redshift radio - galaxies may provide significant numbers of galaxies from which to study the early stages of cluster formation and galaxy evolution .
momentum distributions of nucleons and nucleon clusters in nuclei provide useful insights into various reactions on nuclei , such as @xmath17 and @xmath18 electrodisintegration processes or neutrino - nucleus interaction experiments . variational monte carlo ( vmc ) calculations of these distributions in @xmath19=3,4 nuclei were first reported in @xcite . they were subsequently improved upon and extended to light @xmath20-shell nuclei , and also to the study of nucleon - pair momentum distributions @xcite . many results of these studies have been supplied on request to various researchers , but not formally published before now . hence , in this paper , we detail their current status , provide figures illustrating the features of these momentum distributions , and announce a web site with detailed tabulations available for download @xcite . specifically , we present single - nucleon momentum distributions for @xmath19=212 nuclei , and @xmath21 , @xmath22 , and @xmath23-cluster distributions for @xmath24 nuclei . we also provide a number of nucleon - pair momentum distributions as a function of either the relative pair momentum @xmath25 or the pair center - of - mass momentum @xmath26 . these are projected into spin @xmath27 or 1 and isospin @xmath28 or 1 combinations , or alternatively into @xmath3 , @xmath4 , and @xmath5 pairs . finally we provide more complete calculations of nucleon - pair distributions as a function of both @xmath25 and @xmath26 in @xmath19=3,4 nuclei than previously available . this work is complementary to a major recent study of nucleon momentum distributions by alvioli _ et al . _ they made _ ab initio _ calculations of the @xmath19=2,3,4 @xmath29-shell nuclei using a variety of realistic two - nucleon interactions , as well as more approximate calculations of the larger closed - shell nuclei @xmath30o and @xmath31ca , with an emphasis on the spin - isospin origin of the distributions . our work fills in the light @xmath20-shell nuclei up to @xmath32c , using primarily ( for reasons discussed below ) a realistic two- plus three - nucleon interaction . where our calculations overlap , we are in substantial quantitative agreement with their results . the vmc wave functions used in our calculations have been obtained for a realistic hamiltonian consisting of nonrelativistic nucleon kinetic energy , the argonne @xmath33 ( av18 ) two - nucleon ( @xmath34 ) @xcite , and urbana x ( ux ) three - nucleon ( @xmath35 ) interactions . urbana x is intermediate between the urbana ix ( uix ) and illinois-7 ( il7 ) potentials @xcite . it has the long - range two - pion @xmath36-wave term and short - range central repulsion of uix supplemented with a two - pion @xmath37-wave term , with the strengths of these three terms taken from il7 @xcite ; it does not have the three - pion - ring or isospin - dependent short - range repulsive terms of il7 . the vmc wave functions have the general form @xmath38 \|\psi_j\rangle \ , \ ] ] where @xmath39 is a symmetrization operator and @xmath40 and @xmath41 are two- and three - body correlation operators . the single - particle jastrow wave function @xmath42 is fully antisymmetric and has the @xmath43 quantum numbers of the state of interest . this construction has been described in detail elsewhere @xcite and the variational parameters in @xmath44 have been optimized by minimizing the variational energy expectation value , subject to the constraint of obtaining approximately correct charge radii . most of these wave functions have been used as starting trial functions for recent calculations of energies , electromagnetic moments and transitions , and spectroscopic overlaps using the more accurate green s function monte carlo ( gfmc ) method @xcite . the probability of finding a nucleon with momentum @xmath45 and spin - isospin projection @xmath46,@xmath47 in a given nuclear state is proportional to the density @xmath48 @xmath49 is the spin - isospin projection operator for nucleon @xmath50 , and @xmath51 is the nuclear wave function with total spin @xmath52 and spin projection @xmath53 . the normalization is @xmath54 where @xmath55 is the number of spin - up or spin - down protons or neutrons . the fourier transform in eq . ( [ eq : rhok ] ) is computed by monte carlo ( mc ) integration . a standard metropolis walk , guided by @xmath56 , is used to sample configurations @xcite . we average over all particles @xmath50 in each configuration , and for each particle , a grid of gauss - legendre points @xmath57 is used to compute the fourier transform . instead of just moving the position @xmath58 in the left - hand wave function away from a fixed position @xmath59 in the right - hand wave function , both positions are moved symmetrically away from @xmath59 , so eq . ( [ eq : rhok ] ) becomes @xmath60 here the polar angle @xmath61 is also sampled by mc integration , with a randomly chosen direction for each particle in each mc configuration . this procedure is similar to that adopted in studies of the nucleon - pair momentum distribution @xcite and has the advantage of very substantially reducing the statistical errors originating from the rapidly oscillating nature of the integrand for large values of @xmath45 . to reach momenta @xmath62 @xmath63 in @xmath10he with good statistics requires integrating to @xmath64=20 fm using 200 gauss - legendre points . the proton momentum distribution @xmath65 in the deuteron is shown in fig . [ fig : deut ] . ( for @xmath66=0 , our wave functions have @xmath67=@xmath68 . ) in this case @xmath65 has been evaluated by direct numerical solution of the schrdinger equation , although we have checked that our mc code gives the same results within statistical errors . the separate contributions of the @xmath37- and @xmath69-wave components of the deuteron wave function are also shown . the @xmath37-wave momentum density has a prominent node at 2 @xmath63 , but this is filled in by the @xmath69-wave momentum density , so the total @xmath65 has a distinctive change of slope at 1.5 @xmath63 , followed by a broad shoulder out to 7 @xmath63 before the first @xmath69-wave and second @xmath37-wave nodes occur . the @xmath69-wave component is due to the pion - exchange tensor force in av18 . the broad shoulder is the dominant feature in all the single - nucleon momentum distributions of larger nuclei shown below . the polarized proton and neutron densities in @xmath7he in the @xmath70 state are shown in fig . [ fig : he3pol ] . the spin - up proton and neutron densities are very similar . the spin - down proton density is slightly larger , particularly in the dip region around 2 @xmath63 . although the spin - down neutron density is non - zero only by virtue of the tensor force , it exceeds the spin - up neutron density in the dip region . the total normalizations for @xmath7he are given in table [ tab : sigma - tau ] , along with those of other nuclei in our study . the total number of spin - down neutrons is 2/3 of the @xmath7he @xmath69-state percentage . these momentum distributions may be instructive when considering the use of polarized @xmath7he as a polarized neutron target . l d d d d nucleus & & & & + @xmath7he(@xmath71 ) & 0.974 & 1.026 & 0.938 & 0.062 + @xmath12li(@xmath72 ) & 1.924 & 1.076 & 1.924 & 1.076 + @xmath14li(@xmath73 ) & 1.934 & 1.066 & 1.981 & 2.019 + @xmath16li(@xmath74 ) & 1.914 & 1.086 & 2.855 & 2.145 + @xmath75li(@xmath73 ) & 1.907 & 1.093 & 3.084 & 2.916 + @xmath75be(@xmath73 ) & 1.994 & 2.006 & 2.880 & 2.120 + @xmath76b(@xmath77 ) & 2.901 & 2.099 & 2.901 & 2.099 [ tab : sigma - tau ] the proton momentum density in @xmath10he is shown in fig . [ fig : he4 ] out to 10 @xmath63 on the same scale as the deuteron @xmath65 in fig . [ fig : deut ] . the overall shape is rather similar to the deuteron , with a change of slope near 2 @xmath63 and a broad shoulder out to a second dip near 8 @xmath63 . the @xmath65 is shown for the standard av18+ux hamiltonian used in this work , and for av18 alone using three different versions of the vmc wave function of eq.([eq : psiv ] ) . the full av18+ux hamiltonian gives a vmc energy of @xmath78 mev ( @xmath79 mev in gfmc ) and a radius of 1.44 fm , which are close to the experimental energy of @xmath79 mev and point proton radius of 1.46 fm . the av18 two - nucleon potential alone gives less binding with a vmc energy of @xmath80 mev ( @xmath81 mev in gfmc ) and a larger point proton radius @xmath82 of 1.52 fm . because av18+ux produces energies and radii that are closer to experiment for all @xmath83 nuclei , this is the primary model we use for this paper . the main role of the three - nucleon potential is to fix both the binding energy and size of the nuclei , which for the @xmath65 translates into an overall shift of the momentum density to larger @xmath45 , with a reduced value for @xmath84 @xmath63 . the @xmath10he @xmath69-state increases from 13% with av18 alone to 15% , and the @xmath65 beyond 2 @xmath63 increases by 10 - 20% . to study the source of the intermediate- and high - momentum densities , we show two av18 calculations with simplified wave functions : 1 ) a pure central jastrow wave function labeled av18(j ) in the figure , and 2 ) a wave function including spin - isospin operator correlations , but no tensor components , labeled av18(4 ) . the av18(j ) calculation has just the @xmath85 in eq.([eq : psiv ] ) . the av18(4 ) adds to the central jastrow term the @xmath40 pair correlation operator with spin , isospin , and spin - isospin components . the full av18 wave function also includes tensor and tensor - isospin components in the @xmath40 operator . ( for av18 alone , there is no @xmath41 correlation . ) the central jastrow term dominates the momentum density for @xmath84 @xmath63 , goes through a minimum at @xmath86 @xmath63 , and then gives about 30% of the total density for @xmath87 @xmath63 . the spin - isospin correlations in the av18(4 ) calculation shift the jastrow result to slightly higher momenta , and then begin to dominate at higher momenta , providing an additional 40 - 60% of the density beyond @xmath88 @xmath63 . the tensor components in the full av18 wave function fill in the minimum in the region @xmath89 @xmath63 , just as they do in the deuteron , but are then only 10 - 20% of the total at higher momenta . the proton momentum distribution for @xmath10he is also shown in closeup in fig . [ fig : he4b ] along with the nucleon - cluster distributions for @xmath8 and @xmath9 overlaps . the definition for these distributions is given in ref . @xcite ; essentially , they are the fourier transforms of spectroscopic overlaps as discussed in ref . the @xmath90 at low momenta is almost on top of @xmath65 , indicating that in this region , most of the residual nucleus is in the triton ground state . the total @xmath91 , compared to @xmath92 , in good agreement with the spectroscopic factor determined in gfmc calculations @xcite . the @xmath93 integrates to @xmath94 , with 0.98 coming from having the @xmath9 pair in a relative @xmath37-wave , and 0.025 from a relative @xmath69-wave . because there are two deuterons in a @xmath9 pair , this means there are two deuterons among the six pairs in @xmath10he from this configuration . ( approximately 0.4 more deuterons should be present in @xmath10he from @xmath95 configurations where the @xmath4 pair is in a @xmath96=0 state orthogonal to the deuteron @xcite . ) the proton and neutron momentum distributions in @xmath12he are shown in fig . [ fig : he6 ] . the proton momentum distribution is close to that in @xmath10he . the two neutrons in the @xmath20-shell make the neutron momentum distribution peak at finite @xmath97 @xmath63 . also shown is the difference @xmath98 , which should give a good approximation to the distribution of the @xmath20-shell halo neutrons . the difference has a dip at zero momentum , appropriate for the @xmath20-shell , and because the halo neutrons are primarily in a relative @xmath99 pair , there is a dip at the usual @xmath37-wave node position . although there are twice as many neutrons as protons in @xmath12he , the high - momentum densities are almost equal ; for @xmath100 @xmath63 the ratio @xmath67/@xmath68 = 1.1 . a similar situation holds for @xmath16he ( not shown here , but illustrated and tabulated online @xcite ) where there are three times as many neutrons as protons , but above 2 @xmath63 the ratio @xmath67/@xmath68 = 1.2 . this supports the recent suggestion of sargsian @xcite that in neutron - rich systems , the fraction of protons at high momenta is substantially larger than the fraction of neutrons . the @xmath12li @xmath101 ground state single - nucleon momentum distribution and the @xmath11 cluster distributions in the ground state and first @xmath77 excited state are shown in fig . [ fig : li6 ] . in this case we plot @xmath102=@xmath68+@xmath67=@xmath103 for comparison to the ground state @xmath11 distribution to illustrate that at zero momentum the two are almost equal . the @xmath104 cluster distribution in the @xmath72 ground state has a node at @xmath105 @xmath63 because the @xmath23 and @xmath21 are in a relative @xmath106 spatial state due to antisymmetry @xcite . the integrated @xmath107 is a sum of @xmath37- and @xmath69-wave parts of 0.846 and 0.017 , respectively . the @xmath104 in the @xmath77 state has the @xmath23 and @xmath21 in a relative @xmath108 spatial state and a somewhat greater @xmath109 . however , @xmath110 for the excited state would be indistinguishable from the ground state in this figure , so it is not shown here , but it is tabulated online . the ground state single - nucleon momentum distributions in @xmath14li are shown in fig . [ fig : li7 ] . the proton momentum distribution is similar to that in @xmath12li ; the neutron distribution is bigger at small finite momentum due to the extra @xmath20-shell neutron . the @xmath13 cluster distribution is shown for both the @xmath73 ground state and the @xmath111 second excited state . the former has a node at zero momentum and a second node at @xmath112 @xmath63 , indicative of the @xmath23 and @xmath22 clusters being in a relative @xmath113 spatial state , with @xmath114 . the @xmath115 first excited state has a similar @xmath116 shape and is not shown . the @xmath111 second excited state has the @xmath23 and @xmath22 clusters in a relative @xmath117 spatial state ; the @xmath118 third excited state ( not shown ) has a similar cluster structure . the single - nucleon and @xmath13 cluster distributions for all four states are given online @xcite . the proton momentum distribution for the @xmath119 ground state of @xmath16be is shown in fig . [ fig : be8 ] along with the @xmath15 cluster distributions in both the @xmath119 and first excited @xmath74 states . the @xmath65 for the @xmath74 state at 3 mev excitation , which is not shown , is almost identical with the ground state , varying by @xmath120 to @xmath121 out to 5 @xmath63 . at very low momenta , the nucleus can be considered to be primarily in an @xmath15 cluster , but this probability declines rapidly as @xmath45 increases . the @xmath122 for the @xmath119 state has two nodes because the @xmath23 s are in a relative @xmath123 spatial state , but for the @xmath74 rotational excited state they are in a relative @xmath124 spatial state . the integrated @xmath125 are 0.84 and 0.83 , respectively . the next state in the @xmath16be spectrum is the @xmath126 rotational state at 11 mev excitation ; its @xmath65 varies from the ground state by @xmath127 to @xmath128 over the momentum range shown here , while its @xmath122 exhibits the structure of a @xmath129 spatial state , with a smaller @xmath130 . the @xmath65 and @xmath122 for all three states are tabulated online . the proton and neutron momentum distributions in @xmath76be are shown in fig . [ fig : be10 ] . the proton distribution is very much like that in @xmath16be , while the two added neutrons are mostly at small , but finite @xmath45 as appropriate for the @xmath20-shell . also shown is the difference @xmath98 ; this difference is very much like the two halo neutrons in @xmath12he shown in fig . [ fig : he6 ] , again with a dip at zero momentum , and with a dip at the usual @xmath37-wave node position , although it is somewhat smeared out compared to the former case . tables and figures of the nucleon momentum distributions are provided online at http://www.phy.anl.gov / theory / research / momenta/. the tables give the proton momentum distribution , with one standard deviation error bars , in @xmath96=0 nuclei with the understanding that neutron and proton momentum distributions are identical . for @xmath131 nuclei , both proton and neutron momentum distributions are tabulated . similarly , for @xmath132 nuclei , only total nucleon momentum distributions are given , but for @xmath133 , the spin - up and spin - down distributions for @xmath134 states are also given . in addition , the total normalization @xmath135 for each momentum distribution is given , as are the contributions to the kinetic energy . the corresponding spatial densities for these wave functions are provided online at http://www.phy.anl.gov / theory / research / density/. included in the tables , but not shown here , are distributions for the ground states of @xmath7h , @xmath16he , @xmath136li , @xmath75be , and @xmath76b . we have also made preliminary calculations of the nucleon momentum distributions in @xmath137b and @xmath32c . in these cases , the jastrow part of the variational wave function has been made with only the most spatially symmetric components , [ 443 ] and [ 444 ] , respectively . the web site will be updated when calculations are made with more sophisticated wave functions for these and other nuclei . the overall evolution of the proton momentum distribution in the @xmath96=0 nuclei is illustrated in fig . [ fig : allt0 ] . the shape of the distributions shows a smooth progression as nucleons are added . as @xmath19 increases , the nuclei become more tightly bound , and the fraction of nucleons at zero momentum decreases . as nucleons are added to the @xmath20-shell , the distribution at low momenta becomes broader , and develops a peak at finite @xmath45 . the sharp change in slope near @xmath86 @xmath63 to a broad shoulder is present in all these nuclei and is attributable to the strong tensor correlation induced by the pion - exchange part of the @xmath34 potential , boosted by the two - pion - exchange part of the @xmath35 potential . above @xmath138 @xmath63 , the bulk of the momentum density appears to come from short - range spin - isospin correlations . the probability of finding two nucleons in a nucleus with relative momentum @xmath139 and total center - of - mass momentum @xmath140 in a given spin - isospin state is given by : @xmath141 where @xmath142 , @xmath143 , and @xmath144 is a projector onto pair spin @xmath27 or 1 , and isospin @xmath28 or 1 . the total normalization is : @xmath145 where @xmath146 is the total number of nucleon pairs with given spin - isospin .. alternate projectors can also be used , e.g. , for @xmath34 pairs @xmath3 , @xmath147 , and @xmath5 with corresponding normalizations . the nucleon - pair momentum distributions can be examined in a number of different ways . one way is to integrate over all values of @xmath148 and reduce the pair density to a function @xmath149 of the relative momentum @xmath25 only . in this case , eq.([eq : rhoqq ] ) reduces to a form similar to eq.([eq : actual ] ) , with a sum over all configurations in the monte carlo walk controlled by @xmath150 , and a gauss - legendre integration over the relative separation @xmath151 . again , the polar angle @xmath152 is sampled by randomly choosing the direction of @xmath153 in space , and an average over all pairs in every mc configuration is made . typical results are shown in fig . [ fig : he4st ] for @xmath10he and in fig . [ fig : be9st ] for @xmath75be , where both the total distribution and the different @xmath2 components are plotted . the spatial distribution of nucleon pairs in different @xmath2 combinations , using correlated wave functions generated by realistic interactions , was first studied in ref . @xcite for @xmath154 nuclei . the corresponding momentum distributions have been calculated here for a number of cases in addition to those shown in the figures , including @xmath7he , @xmath155li , @xmath16be , @xmath76b , and are available online @xcite . l c d d d d nucleus & @xmath156 & & & & + @xmath7he(@xmath71 ) & ip & 1.5 & 0.0 & 1.5 & 0.0 + & cor & 1.36 & 0.14 & 1.49 & 0.01 + @xmath10he(@xmath119 ) & ip & 3.0 & 0.0 & 3.0 & 0.0 + & cor & 2.54 & 0.46 & 2.99 & 0.01 + & [ av18 ) & 2.60 & 0.40 & 2.99 & 0.01 + & [ av18(4 ) ] & 2.99 & 0.01 & 2.99 & 0.01 + & [ av18(j ) ] & 3.00 & 0.00 & 3.00 & 0.00 + @xmath12li(@xmath72 ) & ip & 4.5 & 4.5 & 5.5 & 0.5 + & cor & 4.05 & 4.95 & 5.47 & 0.53 + @xmath14li(@xmath73 ) & ip & 6.75 & 6.75 & 6.75 & 0.75 + & cor & 6.12 & 7.38 & 6.74 & 0.76 + @xmath16be(@xmath119 ) & ip & 9.0 & 9.0 & 9.0 & 1.0 + & cor & 8.07 & 9.93 & 8.98 & 1.02 + @xmath75be(@xmath73 ) & ip & 10.5 & 13.5 & 10.5 & 1.5 + & cor & 9.56 & 14.44 & 10.48 & 1.52 + @xmath76b(@xmath77 ) & ip & 12.0 & 18.0 & 13.0 & 2.0 + & cor & 11.17 & 18.83 & 12.95 & 2.05 [ tab : st ] the number of @xmath146 pairs for these nuclei are given in table [ tab : st ] . in the absence of spin - isospin correlations , the total number of @xmath2 pairs in a given nucleus depends on the total number of @xmath34 pairs @xmath157 , the total nuclear spin @xmath158 and isospin @xmath159 , and the spatial symmetry of the wave function as described in ref . this is the independent - pair ( ip ) value given in the table . ( for @xmath20-shell nuclei there are multiple spatial symmetry states available ; we quote the ip values for the most symmetric state , which is always the dominant ground - state component for realistic @xmath34 interactions . ) for the @xmath29-shell nuclei , only @xmath2 combinations ( 10 ) and ( 01 ) with even orbital angular momentum are present . in the @xmath20-shell , the @xmath2 combinations ( 11 ) and ( 00 ) with odd orbital angular momentum start to contribute . for @xmath160 nuclei , the ( 11 ) pairs become the most numerous of all ; in the nuclear matter limit , the ratio ( 11):(10):(01):(00 ) is 9:3:3:1 . when spin - isospin correlations , and particularly the tensor - isospin correlations , are switched on , there is a noticeable transmutation of @xmath2 = ( 01 ) pairs into ( 11 ) pairs , relative to the ip values . there is also a much smaller conversion of @xmath2 = ( 10 ) pairs to ( 00 ) pairs . ( because isospin is largely conserved for realistic nuclear forces , we use wave functions of fixed total isospin , so the sum of ( 01 ) and ( 11 ) pairs is unchanged from the ip value , as is the sum of ( 10 ) and ( 00 ) pairs . ) these correlated values are in the rows labeled `` cor '' in table [ tab : st ] . this transmutation comes about because tensor correlations can flip the spin of a nucleon @xcite , e.g. , if a pair of nucleons is in a ( 01 ) combination , a tensor correlation to a third nucleon has an amplitude to flip the spin of one of the pair nucleons and change the pair to ( 11 ) . because realistic @xmath34 interactions are much more repulsive in @xmath2 = ( 00 ) than ( 11 ) combinations , optimized correlations will only permit a much smaller transmutation of ( 10 ) pairs into ( 00 ) pairs . that the tensor correlations are the source of this effect is demonstrated in the case of @xmath10he by calculations using the av18 interaction alone , and the av18(4 ) and av18(j ) truncated wave functions discussed above in conjunction with fig . [ fig : he4 ] ; their @xmath146 values are given in table [ tab : st ] in the rows labeled [ av18 ] , [ av18(4 ) ] , and [ av18(j ) ] . the pure jastrow wave function without any spin - isospin correlations generates no ( 11 ) or ( 00 ) pairs at all . the av18(4 ) wave function without tensor forces generates ( 11 ) and ( 00 ) pairs with only a tiny amplitude of 0.01 . the av18 interaction with a realistic tensor force converts 0.40 ( 01 ) pairs to ( 11 ) pairs . addition of the @xmath35 potential boosts this effect about 15% to 0.46 pairs , proportional to the increase it induces in the @xmath10he @xmath69-state percentage . from the results shown in table [ tab : st ] for @xmath10he and @xmath16be , from our preliminary results for @xmath32c , and from the @xmath30o and @xmath31ca results of ref . @xcite , it appears that @xmath1610.5 ( 10 ) pairs are converted to ( 11 ) pairs for every @xmath23-particle cluster in the nucleus . in the @xmath149 projections for @xmath10he shown in fig . [ fig : he4st ] , the even partial - wave @xmath2 = ( 10 ) and ( 01 ) pairs are seen to have a similar shape , with an @xmath37-wave peak out to @xmath25=1 @xmath63 followed by a change of slope with a broad shoulder extending beyond 5 @xmath63 . at higher momenta , the ( 10 ) pairs dominate , while the ( 01 ) pairs show a relative dip in the @xmath25=1.52 @xmath63 range . the odd partial - wave ( predominantly @xmath36-wave ) ( 11 ) pairs vanish at @xmath162 but grow to equal the ( 10 ) pairs and exceed the ( 01 ) pairs in the dip region . the ( 00 ) pairs also have a @xmath36-wave shape , but are much smaller in magnitude . for @xmath163 @xmath63 there are 0.56 ( 10 ) pairs compared to 0.29 ( 01 ) pairs and 0.29 ( 11 ) pairs . the @xmath149 projections for @xmath75be shown in fig . [ fig : be9st ] have a very similar structure , except that the total number of ( 11 ) pairs has grown significantly ( as expected from the ip count ) to become the largest component for @xmath25=0.52.5 @xmath63 . the number of ( 00 ) pairs has also grown , but remains much smaller than the others . at higher momenta , the ( 10 ) pairs still dominate the tail of the momentum distribution . instead of projecting @xmath149 into @xmath2 pairs , we can project out @xmath3 , @xmath147 , and @xmath5 pairs . an example of this is shown in fig . [ fig : he8qnn ] for @xmath16he . in this neutron - rich nucleus , there are 15 @xmath5 pairs , 12 @xmath147 pairs , and only 1 @xmath3 pair . the excess of @xmath5 over @xmath147 pairs is all at low relative momenta @xmath164 @xmath63 , where 13.5 ( 10.2 ) @xmath5 ( @xmath147 ) pairs reside . the total number of pairs with @xmath100 @xmath63 is small , but there are 70% more @xmath147 than @xmath5 pairs in this regime . calculations of @xmath165 for @xmath166he , @xmath12li , @xmath16be , and @xmath76b are also provided online @xcite . the @xmath167 in the helium isotopes characterizes the @xmath29-shell core and shows only small changes as neutrons are added to go from @xmath10he to @xmath12he and @xmath16he . this is consistent with the findings of ref . @xcite where only small differences in the two - proton spatial densities were found , indicating the @xmath23 core in @xmath168he is fairly stiff against deformation . we can also integrate eq.([eq : rhoqq ] ) over all @xmath169 , leaving a function @xmath170 of the total pair momentum @xmath26 only . in fig . [ fig : he8qnn ] we show @xmath170 projected into @xmath34 pairs , again for @xmath16he . in general , the @xmath171 for a given nucleus has a smaller falloff at large momenta than the @xmath165 and the ratios of different @xmath34 components vary less over the range of @xmath26 . in addition to @xmath16he , calculations of @xmath171 for @xmath172he , @xmath12li , and @xmath16be are provided online @xcite . a simple internal consistency check on our integration methods can be made using @xmath7he . by virtue of momentum conservation , when a @xmath3 pair has total momentum @xmath148 , the remaining neutron has momentum @xmath173 . consequently , @xmath174 for @xmath175 and similarly @xmath176 . this relation is satisfied to better than 1% as seen by comparing the relevant online tables . we can also again contrast the @xmath177 among the @xmath178he isotopes ; they are similar in shape , but show a noticeably greater variation than the @xmath179 because the center - of - mass is no longer at the center of the @xmath23 core once the @xmath20-shell neutrons are added . the @xmath180 are ordered proportional to the charge radius , i.e. , the value for @xmath12he is largest and that for @xmath10he is smallest . nucleon - pair momentum distributions @xmath181 as a function of both @xmath169 and @xmath148 are more expensive to evaluate as eq.([eq : rhoqq ] ) then requires a double gauss - legendre integration for each pair in each mc configuration ; see eq.(3 ) of ref . nucleon - pair distributions for the special case @xmath26=0 were evaluated for a number of @xmath24 nuclei in @xcite , to help understand the results of a two - nucleon knockout experiment on @xmath32c at jefferson lab @xcite . the experiment showed a ratio of back - to - back @xmath147 to @xmath3 pairs of nearly 20 ( compared to the simple total ratio of 2.4 ) in the relative momentum range @xmath25=1.53 @xmath63 . this is easily understood by the presence of a noticeable dip in the @xmath26=0 @xmath3 distribution for this range of @xmath25 relative to the @xmath147 distribution , as shown in fig . [ fig : be8qq ] for @xmath16be . in addition to this case , our latest results at @xmath26=0 for @xmath182he , @xmath155li , and @xmath76b are provided in the online tables @xcite . unfortunately , @xmath32c is currently still too expensive to evaluate with this double integral , but we are exploring more efficient ways of evaluating eq.([eq : rhoqq ] ) that should make it possible in the future . nucleon - pair distributions for @xmath7he and @xmath10he with finite @xmath183 were reported in @xcite . we now have made much more extensive calculations for these two nuclei averaged over all directions of @xmath148 and @xmath169 . the @xmath184 for @xmath10he is shown in fig . [ fig : pphe4 ] and the @xmath185 is shown in fig . [ fig : pnhe4 ] as a series of @xmath25-dependent curves for fixed values of @xmath26 from 0 to 1.25 @xmath63 . the trends illustrated by these figures are similar to the single - nucleon momentum distributions . the @xmath3 pairs are primarily in relative @xmath186s@xmath187 states , and exhibit the usual @xmath37-wave node for total pair momenta @xmath26=0 . as @xmath26 increases , this node begins to fill in , until it is only a dip for @xmath188 @xmath63 . the @xmath4 pairs are predominantly in deuteron - like states , with the @xmath69-wave contribution filling in the @xmath37-wave node and beginning the usual broad shoulder above 2 @xmath63 . the overall magnitude of the curves for both @xmath3 and @xmath4 pairs decreases as @xmath26 increases , simply because there are fewer pairs with high total momenta . the numerical values for these curves may be found online along with similar results for @xmath7he @xcite . additional calculations for various fixed angles between @xmath148 and @xmath169 are in progress . we have calculated a large number of nucleon , nucleon - cluster , and nucleon - pair momentum distributions for @xmath0 nuclei using vmc wave functions generated from a realistic hamiltonian containing the av18 @xmath34 and ux @xmath35 potentials . the single - nucleon @xmath189 have a common characteristic shape , with a peak at zero ( for @xmath29-shell nuclei ) or small ( for @xmath20-shell nuclei ) momentum , a rapid drop of more than two orders of magnitude followed by a distinct change of slope around @xmath25=1.52 @xmath63 , and then an extended high - momentum tail that carries well beyond 5 @xmath63 . the dominant source of this tail in the 1.53 @xmath63 range is the @xmath34 tensor force , which comes from the pion - exchange potential , and is boosted @xmath16115% by the @xmath35 force with its two - pion - exchange terms . spin - isospin correlations appear to dominate at higher momenta , and presumably are dependent on the short - range structure of the hamiltonian . our calculations include a breakdown between spin - up and spin - down nucleons in @xmath133 nuclei and between protons and neutrons in @xmath131 nuclei . nucleon - cluster distributions , with @xmath21 , @xmath22 , and @xmath23 clusters , have been calculated in @xmath24 nuclei , including a number of excited states . they do not exhibit high - momentum tails but have specific nodal structures that reflect the requirements of antisymmetry for the given spin @xmath52 of the nuclear state . the nucleon - pair momentum distributions @xmath149 as a function of relative pair momentum @xmath25 have a very similar structure to @xmath189 . we provide the breakdown into pair spin - isospin @xmath2 channels in a number of cases , and alternatively into @xmath34 channels in other cases . these different components have different behaviors as a function of @xmath25 which can again be traced back primarily to the pion - exchange tensor force , with the high - momentum tails largely dominated by the @xmath2 = ( 10 ) or @xmath147 pairs . the distributions @xmath170 as a function of total pair momentum @xmath26 also have a similar shape , but are noticeably flatter and seem to have less interesting structure . the @xmath181 are more complicated , particularly with the evolution of nodal structures as illustrated in fig.[fig : pphe4 ] . while we have illustrated many of the above cases in the figures of this paper , the main results are to be found in our online figures and tabulations @xcite . these include single - nucleon momentum distributions for 16 different nuclear ground states plus some excited states . the corresponding configuration - space densities are also provided . nucleon - cluster distributions are given for 12 different cases . nucleon - pair momentum distributions include eight @xmath190 , seven @xmath165 , six @xmath171 , and six @xmath191 nuclear ground states , plus two dozen @xmath192 cases in @xmath182he . it is our intention to add additional calculations as they are developed and requests for specific cases will be entertained . in the future it may be possible to evaluate the momentum distributions using the more accurate gfmc wave functions @xcite . in that case , eq.([eq : rhok ] ) will have to be evaluated moving only @xmath193 , which is an inherently noisier procedure . the most likely result from a gfmc evaluation is a further shift to higher momenta . first , the gfmc wave functions produce more binding which will decrease somewhat the population of nucleons at low momenta . second , gfmc wave functions generally have spin - isospin and tensor correlations that are enhanced relative to the vmc wave functions , which will raise the broad shoulder in the 2 to 7 @xmath63 region ; these correlations should also enhance polarization . the gfmc kinetic energy , i.e. , the second moment of the nucleon momentum distribution , is generally higher than the vmc kinetic energy by @xmath194 , also indicating a shift of nucleons to higher momenta , but still subject to the overall normalization remaining the same . this work is supported by the u.s . department of energy , office of nuclear physics , under contracts de - ac02 - 06ch11357 ( rbw and scp ) , de - ac05 - 06or23177 ( rs ) , de - ac52 - 06na25396 ( jc ) , and the nuclei scidac grant . the calculations were made on the parallel computers of argonne s laboratory computing resource center . 100 r. schiavilla , v. r. pandharipande , and r. b. wiringa , nucl . phys . * a449 * , 219 ( 1986 ) . r. schiavilla , r. b. wiringa , s. c. pieper , and j. carlson , phys . rev . lett . * 98 * , 132501 ( 2007 ) . r. b. wiringa , r. schiavilla , s. c. pieper , and j. carlson , phys . rev , c * 78 * , 021001(r ) ( 2008 ) . see http://www.phy.anl.gov/theory/research/momenta/ for single - nucleon and nucleon - cluster distributions and http://www.phy.anl.gov/theory/research/momenta2/ for nucleon - pair distributions ; corresponding single - nucleon spatial densities are available at http://www.phy.anl.gov/theory/research/density/ . m. alvioli , c. ciofi degli atti , l. p. kaptari , c. b. mezzetti , and h. morita , phys . rev , c * 87 * , 034603 ( 2013 ) . r. b. wiringa , v. g. j. stoks , and r. schiavilla , phys . c * 51 * , 38 ( 1995 ) . s. c. pieper , v. r. pandharipande , r. b. wiringa , and j. carlson , phys . c * 64 * , 014001 ( 2001 ) . s. c. pieper , aip conf . * 1011 * , 143 ( 2008 ) . r. b. wiringa , phys . c * 43 * , 1585 ( 1991 ) . b. s. pudliner , v. r. pandharipande , j. carlson , s. c. pieper , and r. b. wiringa , phys . c * 56 * , 1720 ( 1997 ) . i. brida , s. c. pieper , and r. b. wiringa , phys . c * 84 * , 024319 ( 2011 ) . e. a. mccutchan , c. j. lister , s. c. pieper , r. b. wiringa , d. seweryniak , j. p. greene , p. f. bertone , m. p. carpenter , c. j. chiara , g. gurdal , c. r. hoffman , r. v. f. janssens , t. l. khoo , t. lauritsen , and s. zhu , phys . rev . c * 86 * , 014312 ( 2012 ) . s. pastore , s. c. pieper , r. schiavilla , and r. b. wiringa , phys . c * 87 * , 035503 ( 2013 ) . v. m. datar , _ et al . lett . * 111 * , 062502 ( 2013 ) . m. m. sargsian , arxiv:1210.3280v2 ( 2013 ) . j. l. forest , v. r. pandharipande , s. c. pieper , r. b. wiringa , r. schiavilla , and a. arriaga , phys . c * 54 * , 646 ( 1996 ) . r. b. wiringa , phys . c * 73 * , 034317 ( 2006 ) . r. subedi _ et al . _ , science * 320 * , 1476 ( 2008 ) . r. b. wiringa , s. c. pieper , j. carlson , and v. r. pandharipande , phys . c * 62 * , 014001 ( 2000 ) .	we report variational monte carlo calculations of single - nucleon momentum distributions and nucleon - pair and nucleon - cluster momentum distributions for @xmath0 nuclei . the wave functions have been generated for a hamiltonian containing the argonne v@xmath1 two - nucleon and urbana x three - nucleon potentials . the single - nucleon and nucleon - pair momentum distributions exhibit universal features attributable to the pion - exchange tensor interaction . the single - nucleon distributions are broken down into proton and neutron components and spin - up and spin - down components where appropriate . the nucleon - pair momentum distributions are given either in pair spin and isospin @xmath2 projection or for @xmath3 , @xmath4 , and @xmath5 pairs . the nucleon - cluster momentum distributions include @xmath6 in @xmath7he , @xmath8 and @xmath9 in @xmath10he , @xmath11 in @xmath12li , @xmath13 in @xmath14li , and @xmath15 in @xmath16be . detailed tables are provided online for download .
on september 2015 , the gravitational - wave ( gw ) was directly detected by the advanced laser interferometer gravitational - wave observatory ( aligo ) for the first time @xcite . the detected gw , labeled by gw150914 , was the signal from the inspiral and merger of two black holes with the masses of about 30 @xmath2 . the signal was the first evidence of the binary black hole merger . as another event of black hole merger , gw151226 , had been detected @xcite , further details of binary black hole systems , such as distribution and population of black holes , might be revealed by further observations . while aligo detected the transients with the durations of a few hundreds milliseconds , a longer observations of the binary systems would provide much information of the binary systems , such as spins of black holes . binary systems with the masses of about 30 @xmath2 emit gws with the frequencies of about 0.1 hz at 15 days before the merger . such a long term observation of the gw signals from the binary systems would effectively solve the degeneracy of the parameters , such as their spins , which is expected to be a clue about the evolution of the black holes @xcite . in addition to the black holes with the masses of the order of 10 @xmath2 , it is important to search for black holes with various masses . observation of various black hole mergers would elucidate the evolution process of super - massive black holes . in order to observe the gw signals from binary systems with various masses , it is necessary to observe gws at various frequencies . binaries with heavier masses emit gws at lower frequencies than the observation band of the interferometers . a torsion - bar antenna ( toba ) @xcite is a gravitaional - wave ( gw ) detector that is sensitive to gws at around 1 hz , while the observation band of interferometric gw detectors , such as ligo , are above about 10 hz . one of the astronomical targets of toba is intermediate - mass black hole binaries . the space - borne interferometric detectors , such as laser interferometer space antenna ( lisa ) @xcite and deci - herz gravitational - wave observatory ( decigo ) @xcite , would also have the observation frequency band below 10 hz . comparing to such detectors , toba can be realized on the ground with an accessibility for repair and upgrade , and without risk for large cost . though the final target sensitivity of toba is not as good as space detectors , the observation range is expected to be as far as 10 gpc with the luminosity distance for the intermediate - mass black hole mergers . toba has two bar - shaped test masses which are suspended at their centers . gws are detected by monitoring their relative rotation excited by the tidal force from gws . since the resonant frequencies of the test masses in the torsional modes are as low as a few mhz , toba has sensitivity at low frequencies even on the ground . as a proof of concept , the first prototype of toba had been developed with a single test mass bar @xcite and set the first upper limit on a stochastic gw background from 0.03 to 0.8 hz @xcite . on the other hand , it is necessary to use more than three detectors in order to determine the parameters of binaries , such as masses of the objects , polarization angle , and the sky position of the source because the conventional detectors have poor directivity @xcite . therefore , we proposed _ the multi - output toba _ , which provides three independent signals from a single detector by monitoring multiple rotational degrees of freedom of the test masses @xcite . the multi - output system improves the event rate and the angular resolution , which would enhance the low - frequency gw astronomy even with fewer detectors . in this paper , we introduce the first multi - output toba detector . its main feature is the new suspension system for the multi - output configuration , which also performs as a vibration isolation system . in the section [ sec : toba ] , we explain its principle and the target sensitivity . the detector configuration that we developed is described in the section [ sec : detector ] . its characteristics is mentioned in the following section . toba has two bar - shaped test masses that rotate differentially by the tidal force from gws as shown in fig . [ fig : toba ] @xcite . schematic view of the toba test masses.,width=264 ] the equation of the motion of the test mass bar in the rotational mode is @xmath3 where @xmath4 and @xmath5 are angular fluctuation , the moment of inertia , the damping constant , and the spring constant around @xmath6-axis , respectively . @xmath7 is the amplitude of the gw , and @xmath8 is the dynamical quadrupole moment of the test mass for the rotation along the @xmath6 axis . the frequency response of the angular fluctuation of the test mass is derived from eq . ( [ eq : eom ] ) as @xmath9 where @xmath10 and @xmath11 are the loss angle and the amplitude of the gw coming along @xmath6-axis with the polarizations of @xmath12 and @xmath13 , respectively . @xmath14 is the resonant frequency of the torsional mode , above which the angular fluctuation due to the gw is approximated to be independent from its frequency . in the multi - output system , we also consider rotations of the bars on the vertical planes . considering the test mass 1 in fig . [ fig : toba ] , the angular fluctuation along the @xmath15-axis also obeys the similar equation as eq . ( [ eq : eom ] ) : @xmath16 it means that the bar also rotates vertically due to gws coming along @xmath15-axis . therefore , we can derive two independent signals @xmath17 and @xmath18 from the single test mass bar . since we have two orthogonal test mass bars , it is possible to derive three independent signals from the single detector , i.e. , @xmath19 , @xmath20 , and @xmath21 , where the suffix indicates the two test masses . the sensitivity to gw signal derived from @xmath17 , calculated as @xmath22 , would be better by @xmath23 than the sensitivity derived only from one of the rotational signals , when the noise appeared in @xmath24 , and @xmath21 are un - correlated and the same level . this multi - output system would increase the expected detection rate by about 1.7 times , since the three signals have different sensitive areas in the sky . also , parameter estimation accuracy for short - duration signals would be improved since the three independent information helps to break the degeneracies of the parameters @xcite . the final target sensitivity of toba is about @xmath25 @xmath1 in strain at 1 hz as described in @xcite , which is limited mainly by the shot noise , the radiation pressure noise , and gravity gradient noise . this sensitivity can be achieved by sensing the rotation of large cryogenic test mass bars with the length of 10 m , using fabry - perot interferometer . the test mass bars and the suspension wires should be cooled down in order to reduce the thermal noise . since several advanced technologies are necessary for the final toba configuration , it is required to develop each component using prototypes . the first prototype toba had been developed previously @xcite . the first prototype has a single bar shaped test mass with a magnetic levitation system in order to suspend the test mass softly in the rotational degree of freedom . the test mass has the length of 20 cm , and its horizontal rotation is measured by a michelson interferometer , which means that it has the single - output configuration . it successfully tested the basic principle of toba , and set the first upper limit on a stochastic gravitational - wave background at 0.2 hz @xcite . the sensitivity of the first prototype toba is limited by the magnetic noise induced by the magnetic suspension system , and the seismic noise coupling . as a next step , we have developed a multi - output toba , as described in the following sections . its main target is the development of the suspension system that realizes the multi - output system . also , the passive and active vibration isolation systems attenuate the noise caused by the seismic motion . ( color online ) . overview of the multi - output toba.,width=264 ] the schematic view of the multi - output toba is shown in fig . [ fig : schemview ] . the two orthogonal test masses that sense the gws are suspended from a hexapod - type active vibration isolation table ( avit ) via an intermediate mass , which is magnetically damped by a damping mass . the optical bench where the sensors and actuators are set is also suspended from the intermediate mass . the suspension system except the actuators and sensors of the avit is in a vacuum tank . the picture of the test masses is shown in fig . [ fig : testmass ] . the test masses are designed for the multi - output system and the test of common mode noise reduction . ( color online ) . picture of the test masses . each test mass is suspended by two wires so that the center of each mass can be located at the same position.,width=264 ] the two orthogonal test mass bars with the length of 24 cm are suspended by the two parallel wires respectively , so that the centers of the masses can be located at the same position in order to maximize the common mode noise reduction rate in horizontal rotation signal . this design is implemented since it may reduce the noise caused from the common rotational displacements of the bars . the common mode noise reduction rate is expected to be large when the sensitivity is limited by the environmental disturbance that effects the test mass rotations commonly , while the reduction rate is @xmath23 in strain at the minimum when the noise sources of the two signals are independent . the resonant frequency in the horizontal rotational mode is @xmath26 , where @xmath27 , @xmath28 , @xmath29 and @xmath30 are the mass of the test mass , the acceleration of gravity , the distance between the two suspension wires as shown in fig.[fig : testmass ] , moment of inertia of the test mass , and the length of the wire , respectively . the resonant frequency in the vertical rotational mode is written as @xmath31 , where @xmath32 is the height distance between the suspension points and the center of mass as shown in fig.[fig : testmass ] . the suspension points of the test masses are set to be close to their center of mass in order to minimize the resonant frequency of the vertical rotation . in our set up , the resonant frequency of the horizontal and vertical rotational modes are 0.10 hz and 0.15 hz , respectively . the resonant frequencies are set at around 100 mhz in our prototype in order to realize the new suspension system with the compact setup . the resonant frequencies would be pushed down by using larger test masses in the future upgrade in order to widen the observation band . the main sensors that we used for the observation in the multi - output toba are the michelson interferometeric sensors . the motion of the bar is monitored by measuring the phase shift of the beam reflected by the mirrors attached at the bar . since it is necessary to have several sensors around the test masses in order to monitor three independent rotational signals , the fiber optics are used for the space saving . the sensor configuration is shown in fig . [ fig : sensor ] . the type-1 interferometers , which measure the displacement of each mirror attached on the test masses , monitor the yaw , longitudinal , and side motion of the bar . the type-2 interferometers that sense the differential displacement of the two end mirrors are set as sensors for the roll motion . the position of the test masses are controlled by the coil - magnet actuators so that the fringe of the interferometer can be kept at their middle . the gw signals are derived from the feedback signal on the actuators . the test masses are controlled in longitudinal , side , yaw , and roll modes . these sensors are set on the suspended optical table explained in the next sub - section . ( color online ) . configuration of the fiber laser interferometers.,width=264 ] though the seismic motion in the rotational degree of freedom is small , the seismic vibration isolation system is necessary since the translational vibration couples to the rotational signals . for example , non - parallel mirrors at the both ends of the test mass induce the translational seismic noise coupling @xcite . the test masses , and the optical table where the sensors and actuators are set , are suspended in order to attenuate the seismic motion above resonant frequencies of those pendulum modes . however , since the resonant frequencies of the pendulum modes are about 1 hz , an additional seismic isolation system for low frequencies are necessary for toba . therefore , we developed the active vibration isolation table ( avit ) for low - frequency vibration isolation and the whole system is suspended from the avit . ( color online ) . schematic view of the hexapod - type active vibration isolation table ( avit ) . the whole suspension system is suspended from the top table . it has six piezoelastric actuators as legs so that they can actuate the top table in six degrees of freedom . the six seismometers are set in order to sense the vibration of the top table.,width=264 ] the avit is a table with six legs composed of piezoelastic elements ( pzts ) as shown in fig . [ fig : hexa ] . the pzts ( p844.30 , products of phisik instrumente ) move the position of the top plate . those pzts have tips with slits at the both ends in order to avoid non - linear effect when they push or pull the table . the vibration of the top table is suppressed by the feedback control using six seismometers ( l-4c geophones , products of sercel ) set on the top plate and the pzts . note that reflective type position sensors are used in order to measure the dc position of the top table relative to the ground since the lack of the sensitivity of seismometers at low frequencies causes drift of the top table . the seismic displacement of the top table is shown in fig . [ fig : seis_hexa ] . the vibration isolation ratio from the ground motion at 1 hz is achieved to be almost 10 times . its performance is mainly limited by the range of the pzts and the resonance of the frame where the avit is sitting . the avit can attenuate the vibration at around 1 hz even with its compact body of 45 cm in diameter , while the passive vibration isolation requires large setup for low - frequency vibration isolation , such as inverted pendulums @xcite . also the avit is effective to attenuate the vibration of the heat link in the cryogenic system for thermal noise reduction that is planned to be implemented in the future upgrade . it is because the rigid structure enables the avit to suppress the vibration induced directly into the vibration attenuated plate , i.e. , the vibration of the heat link attached at the suspension point . seismic isolation performance of the avit in each degrees of freedom . the dashed lines represent the seismic displacement with the avit off , and the solid lines represent one with the avit on . the coordinate is defined in fig . [ fig : schemview].,width=347 ] the angular fluctuations of the test masses are read by the laser interferometers . the angles are calibrated into the gw signal outputs as follows : @xmath33 where @xmath34 @xmath35 are defined as @xmath36 @xmath37 and @xmath38 are the transfer functions from the gw signal to the angular fluctuations of the test masses and the optical bench derived from eq . ( [ eq : h ] ) , respectively . here , @xmath39 are considered to be zeros since @xmath40 when the test masses are suspended along @xmath41-axis and @xmath15-axis . note that the optical bench is also sensitive to gws since it is also suspended from the intermediate mass . in our case , the @xmath42 are @xmath43 above their resonant frequencies , where @xmath44 and @xmath45 . for the optical bench , @xmath46 @xmath47 @xmath48 are not zero because the side view of the optical bench does not show four - fold symmetry , while @xmath49 is derived to be zeros from @xmath50 . therefore , the calibration factor for the @xmath51 and @xmath52 are about 10 times smaller than @xmath53 in our setup as shown in figs . [ fig : calib ] and [ fig : calib2 ] above 2 hz . calibration factor from the horizontal rotational angle to @xmath53 , which is the gw amplitude equivalent signal derived from the horizontal rotation of the bars . above the resonant frequency of the horizontal rotation at 0.1 hz , the calibration factor is constant as derived from eq . ( [ eq : h]).,width=264 ] calibration factors from the vertical rotational angle to @xmath54 , which is the gw amplitude equivalent signals derived from the vertical rotation of the each bar . the peak at 0.15 hz is the resonant frequency of the vertical rotation of the bar . above the the resonant frequency of the optical bench , which is the 1.1 hz with the quality factor of 10 , the calibration factor is decreased because the optical bench also rotates due to the gw.,width=264 ] gw equivalent noise spectra obtained from the multi - output toba compared with the spectrum of the previous detector . the black , dark gray , and light gray lines are the spectra of @xmath55 , and @xmath52 , respectively . the dotted gray line represents the sensitivity of the previous detector @xcite.,width=264 ] the solid lines in fig . [ fig : strain ] are the gw strain equivalent noise spectra obtained from the multi - output toba . the dotted line is the sensitivity of the first prototype . it shows that the sensitivity of @xmath53 that is represented in the black line in fig . [ fig : strain ] is improved by about 100 times at the maximum compared to the first prototype . figure [ fig : sens_limits ] shows the dominant noise sources in the three signals . the sensitivity of the multi - output toba is limited mainly by the interferometer readout noise and the seismic coupling noise . the readout noise , shown in the solid dark gray lines , is estimated from the readout signal measured with the test masses fixed on the optical bench . it is considered to be induced from the fiber optics since the contribution from the other laser noise source , such as the intensity fluctuation and the frequency fluctuation , are lower than the measured readout noise . the light gray lines are the seismic noise estimated from the motion of the avit table measured by the seismometers on the avit and the transfer functions from the seismometers to the laser sensors directly measured by exciting the avit . while the coupling mechanism to @xmath53 is now under investigation , the seismic noises of @xmath51 and @xmath52 are induced because the translational seismic motions excite the vertical rotations of the test masses . since the heights of the suspension point and of the center of mass are different , the translational force on the suspension point applies the torque in the roll direction . the dotted dark gray lines in the two bottom spectra in fig . [ fig : sens_limits ] are the seismic noise calculated from the theoretical transfer functions from the ground to the test masses . the sensitivity curve , the seismic noise estimated from the measured transfer functions , and the theoretical seismic noise fit well in @xmath51 and @xmath52 . in addition to the large seismic coupling , the small calibration factor from the rotation to the gw amplitude , as derived in the previous section , worsen the sensitivity of @xmath51 and @xmath52 by about 1,000 times than @xmath53 . gw equivalent noise spectra and the spectra of the respective noise sources . the left top , left bottom , and right bottom graphs are the sensitivities and the noise sources of @xmath53 , @xmath51 , and @xmath52 , respectively . the black lines are the gw equivalent noise spectra . the dark gray lines show the readout noise measured with the test mass fixed to the optical bench . the light gray blue lines are the seismic noise estimated from the transfer function from the ground to the sensor directly measured using the avit . the dotted dark gray lines are the seismic noise calculated using the theoretical transfer function from the ground to the sensor . the theoretically calculated seismic noise is not plotted in the left top , since the coupling mechanism of seismic noise to @xmath53 is unknown.,width=321 ] effect of the common mode noise rejection . the black and the dark gray lines are the strain sensitivities calculated using single test mass , 1 and 2 , respectively . the dotted light gray line is the sensitivity derived by subtraction of the signals from the two test masses.,width=226 ] the performance of the common mode noise rejection between the two test masses is shown in fig . [ fig : cmr ] . as described in the section [ sec : tm ] , the centers of masses are designed to be at the same position since the noise may be reduced when the sensitivity is limited by the common rotational displacement . however , subtraction of the two signals is not effective for the noise reduction in our case , since there are almost no coherence in two rotational signals . the readout noise is not correlated between the two independent interferometers . also the seismic coupling noise is not correlated , since the seismic motion in @xmath15 direction couples to the rotation signal of the test mass 2 , and the motion in @xmath41 direction couples to the signal of the test mass 1 in fig . [ fig : toba ] . since there are no correlation between the seismic motions in the orthogonal directions , the two rotational signals are not correlated . sensitivities of signals with the avit on and off . the top , left bottom and right bottom graphs represent the sensitivity of @xmath53 ; @xmath51 , and @xmath52 , respectively . the black lines are the sensitivities with the avit on , while the dashed gray lines are the ones with avit off.,width=347 ] figure [ fig : aviteffect ] shows the efficiency of the avit . the dotted gray lines are the sensitivity measured without vibration attenuation using the avit , while solid black lines are the sensitivity measured with the avit working . the sensitivity at around 1 hz largely improved thanks to the avit . at 4 - 10 hz , the noise levels become worse by the avit . it is supposed to be the control noise induced by the avit since the resonance of the support frame disturbs the avit control loops to have enough phase margin . we performed the observational run from 19:50 jst , december 10 , 2014 to 19:50 jst , december 11 , 2014 . the observation system stably continued for more than 24 hours . the spectrograms of the three signals during the observation are shown in fig . [ fig : spectrogram ] . the noise levels are almost the same for 24 hours except for the earthquake which occurred at around 20:00 , december 10 . ( color online ) . spectrograms of the power density of the three signals during the 24-hour observation from 0.1 to 100 hz.,width=347 ] histograms of the normalized power spectral density of @xmath53 , @xmath51 , and @xmath52 at 1 , 2 , and 8 hz . the dotted line at each panel corresponds to a gaussian distribution.,width=347 ] the histograms of the power spectrum density of the obtained data are shown in fig . [ fig : gaussianity ] . the columns represent the signals , @xmath53 , @xmath51 , and @xmath52 , and the lines represent the frequencies of the collected data , 1 hz , 2 hz , and 8 hz , respectively . the sensitivities shown here are normalized by the average sensitivity over the whole observation . the all three signals are distributed according to gaussian distributions . using these data , we performed several gw search analysis : the search for continuous gws , stochastic gw backgrounds , and intermediate mass black hole binaries . continuous gw signals had been searched for at 6 - 7 hz and set an upper limit of @xmath56 on the dimensionless gw strain at 6.84 hz @xcite . also , at 2.58 hz , a new upper limit @xmath57 is set on the energy of the stochastic gw background , where @xmath58 is the gw energy density per logarithmic frequency interval in units of the closure density and @xmath59 is the hubble constant in units of 100 km / s / mpc @xcite . the gws from the binary systems with the mass of 100 @xmath2 , and 200 - 1000 @xmath2 were searched using the matched filtering method , and no signals had been discovered @xcite . the continuous gw signals and gws from the binary systems were searched at around 1 - 10 hz for the first time . the upper limit on the energy of the stochastic gw background has also been updated using the detector with improved sensitivity . the sensitivity of toba was improved mainly due to the passive and active vibration isolation systems . the pendulum suspension system is used as a passive vibration isolation in order to reduce the seismic motion above 1 hz . the vibration isolation system for the sensors , as well as for the test masses , is effective . in addition , the avit are introduced in order to reduce the seismic motion around 1 hz . while its performance was limited by the resonance of the frame where the avit is sitting and the actuation range of the pzts , the seismic motion of the table is reduced by about 10 times from the ground in the translational three degrees of freedom . the combination of those vibration isolation system improves the sensitivity at around 1 - 10 hz by 10 - 100 times compared with the previous prototype toba . also , the three independent signals are successfully obtained simultaneously and stably as the multi - output signals , while the sensitivity of the signals from the vertical rotational motion are worse than that from the horizontal rotational motion . the sensitivity was different among the three signals because of the large coupling from the seismic motion and the cancellation of the gw signals between the test masses and the optical bench . in order to improve the detector performance , careful optimization of the suspension design is necessary . the reinforcement of the frame for the avit is required in order to gain more vibration attenuation ratio . the shape of the optical bench should have four - fold symmetry in order to improve the sensitivities of @xmath51 and @xmath52 so that the optical bench would not react to gws . also , the positions of the centers of masses of the test masses and the optical bench would be adjusted from the outside of the vacuum tank using moving masses and actuators , such as pico - motors , in order to search for the state that minimize the seismic coupling . such modification also helps to investigate the seismic coupling mechanism in the torsion pendulum further . in addition to the suspension design , the readout system is required to be modified in order to improve the sensitivity . having more space on the optical bench would enable us to use the spacial laser beams instead of the fiber optics . for further sensitivity improvement , it is necessary to upgrade several technologies . other than the seismic attenuation system developed in our paper , the low loss suspension system and the cryogenic system are critical path to achieve the final target sensitivity . also , the newly introduced observation method , the multi - output system , is required to be investigated further , though the demonstration was successful . for example , reduction of the suspension thermal noise in the vertical rotational signals would be the one of the important subjects . also , it is necessary for the two vertical rotation signals to improve the detector configuration by the methods as explained in the previous section in order to achieve the same sensitivities as the horizontal rotation signal . however , since the fundamental noise sources for the vertical rotation are the same as the ones for the horizontal rotation , the achievable sensitivities of @xmath51 and @xmath52 are expected to be almost same as @xmath53 . for the midterm upgrade , gw strain equivalent noise of @xmath60 at 0.1 hz , which is the sensitivity such that the gravity gradient effect , so called newtonian noise , could be observed , would be the target . besides the further optimization for the seismic coupling reduction , the key technology for the mid - term upgrade would be the low - loss suspension for the reduction of the thermal noise in the midterm phase . the gravity gradient effect is the noise caused by the gravity perturbation due to the seismic motion , acoustic sound , motion of the object around the detector , and so on @xcite . it is necessary to establish the method of newtonian noise canceling in this phase for the further sensitivity improvement below 1 hz . conversely , precise detection of gravity gradient signals caused by earthquakes would also be applicable for the early alert of large earthquakes . the gravity gradient signal from a large earthquake is considered to be used as the prompt alert of the earthquake since the gravity signal propagates at the speed of light , which is much faster than the seismic waves @xcite . the full - tensor configuration realized by the multi - output system would also be effective in terms of the gravity gradient detection and cancellation @xcite . for the long term upgrade to achieve the final target sensitivity , @xmath25 @xmath1 , the 10-m scale large test mass is necessary . the rotation of the test mass is required to be read by the fabry - perot interferometer in order to reduce the shot noise . also , the cryogenic system to reduce the thermal noise would be critical in order to improve the sensitivity at low frequencies . with this sensitivity , intermediate mass black holes binaries at the luminosity distance of 10 gpc would be observed with the signal to noise ratio of 5 @xcite . such observational results would be a key to reveal the evolution processes of stellar and supermassive black holes , globular clusters , and galaxies . we have developed the new toba detector that employed the new suspension system for the multi - output system . it successfully worked as the multi - output detector , i.e. , the three independent signals was derived simultaneously from the single detector . it demonstrates the new technique to improve the event rate and the parameter estimation accuracy . also , the sensitivity obtained from the horizontal rotational signal is improved from the previous prototype because of the passive and active vibration systems . especially , avit , which is the compact active vibration isolation system realized with the pzt actuators and seismometers , reduced the seismic displacement at lower frequencies than the resonant frequency of the pendulum . avit is also expected to be effective for the vibration isolation for the heat link for the cryogenic system which is planned to be used in the future . those new technologies would be a new step towards the low - frequency gw astronomy . while it is necessary to develop other technologies , such as a cryogenic system and a large system , observation of the binary black hole systems with multi - output toba would provide various astronomical information . this work is supported by grants - in - aid from the japan society for the promotion of science ( jsps ) , jsps fellows grants no . 26.8636 ( k. e. ) , 24.7531 , and 15j11064 ( a. s. ) . this work also is supported by jsps grants - in - aid for scientific research ( kakenhi ) grant no . 24244031 ( m. a. ) , 24103005 , 15h02082 , and 15k05070 ( y. i. ) . 23ifxundefined [ 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 http://link.aps.org/doi/10.1103/physrevlett.116.061102 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.3847/2041 - 8205/818/2/l22 [ * * , ( ) ] link:\doibase 10.1093/ptep / ptw127 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.161101 [ * * , ( ) ] link:\doibase 10.1007/s10714 - 014 - 1730 - 2 [ * * , ( ) ] link:\doibase 10.1088/0264 - 9381/28/9/094011 [ * * , ( ) ] link:\doibase 10.1016/j.physc.2010.05.219 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.161101 [ * * , ( ) ] @noop * * , ( ) http://stacks.iop.org/0264-9381/28/i=12/a=125023 [ * * , ( ) ] link:\doibase 10.1103/physrevd.90.064039 [ * * , ( ) ] _ _ , @noop ph.d . thesis , ( ) _ _ , @noop ph.d . thesis , ( ) link:\doibase 10.1016/j.nima.2007.08.161 [ * * , ( ) ] link:\doibase 10.1016/j.precisioneng.2014.09.010 [ * * , ( ) ] link:\doibase 10.1063/1.1144249 [ * * , ( ) ] link:\doibase 10.1093/ptep / ptv179 [ * * , ( ) ] link:\doibase 10.1103/physrevd.94.042003 [ * * , ( ) ] @noop ( ) link:\doibase 10.1103/physrevd.30.732 [ * * , ( ) ] link:\doibase 10.1093/gji / ggv090 [ * * , ( ) ] @noop * * , ( )	we have developed a new gravitaional - wave ( gw ) detector , torsion - bar antenna ( toba ) , with multiple - output configuration . toba is a detector with bar - shaped test masses that rotate by the tidal force of the gws . in our detector , three independent information about the gw signals can be derived by monitoring multiple rotational degrees of freedom , i.e. , horizontal rotations and vertical rotations of the bars . since the three outputs have different antenna pattern functions , the multi - output system improves the detection rate and the parameter estimation accuracy . it is effective in order to obtain further details of the gw sources , such as population and directions . we successfully operated the multi - output detector continuously for more than 24 hours with stable data quality . also , the sensitivity of one of the signals is improved to be @xmath0 @xmath1 at 3 hz by the combination of the passive and active vibration isolation systems , while sensitivities to possible gw signals derived from the vertical rotations are worse than that from the horizontal rotation .
optical systems are one of the major candidates for implementations of quantum information processing . there are different ways for qubit encoding for optical quantum information processing . probably , the most well - known method is to use a single photon with its polarization degree of freedom . quantum teleportation experiments have been performed using such polarized single photons ( psps ) as qubits @xcite and quantum computing protocols based on linear optics have been developed along this line @xcite . it is also possible to use the vacuum and single - photon ( vsp ) states as the basis for qubit encoding @xcite . coherent - state qubits have been studied as an alternative approach to optical quantum information processing @xcite with their advantages in teleportation @xcite . efficient transfer of qubits is an important factor in quantum information processing . it is particularly crucial for quantum communication and quantum networks @xcite . a comparison among the different types of qubits in terms of transfer efficiencies would be indispensable in order to build an efficient hybrid architecture for optical quantum information processing @xcite in a lossy environment . there are different ways to transfer qubits , for example , such as direct transmission and quantum teleportation @xcite . compared @xcite the teleportation scheme for continuous - variable states @xcite with the direct transmission through a noisy channel . they showed that the teleportation scheme shows better transmission performance than the direct transmission in strong decoherence regions @xcite . park and jeong compared effects of photon losses and detection inefficiency on entangled coherent states and entangled photon - polarized states for quantum teleportation @xcite . extending these investigations , we are interested in comparisons for both the direct transmission and teleportation with the three aforementioned qubit - based approaches . in this paper , we investigate and compare fidelities of information transfer for the three different types of photonic qubits over a lossy environment . we find that teleportation is more robust to photon losses than the direct transmission for vsp qubits , psp qubits , and coherent - state qubits with small amplitudes . while vsp qubits are the most robust ones to photon losses , coherent - state qubits with small amplitudes are more robust than the psp qubits for optical quantum information transfer . in terms of the success probabilities for quantum teleportation based on linear optics , vsp qubits and coherent - state qubits are found to outperform psp qubits under photon loss effects . + we first consider a vsp qubit , @xmath1 , where @xmath2 and @xmath3 are the vacuum and single - photon states , respectively . this type of encoding strategy is sometimes referred to as the single - rail logic because it is defined by the occupation of a single optical mode @xcite . state preperations and operations have been demonstrated experimentally using the single - rail logic @xcite . the time evolution of density operator @xmath4 under photon losses is governed by the born - markov master equation @xcite , @xmath5 where @xmath6 is the interaction time , @xmath7 , @xmath8 , @xmath9 is the decay constant , and @xmath10 ( @xmath11 ) is the annihilation ( creation ) operator for mode @xmath12 . the general solution of eq . is written as , @xmath13\rho(0)$ ] , where @xmath14 is the initial density operator @xcite . a vsp qubit under the direct transmission with photon losses is simply obtained as @xmath15 where @xmath16 , @xmath17 , @xmath18 and @xmath19 . the average fidelity between input and output states is @xmath20 a schematic comparison between the direct transmission and the teleportation process is presented in fig . [ scheme ] . in general , the quantum teleporation protocol for a qubit @xcite requires a bipartite entangled state as the quantum channel in addition to a bell - state measurement scheme that discriminates the four entangled states called the bell states . the sender s outcome for the bell - state measurement is sent to the receiver through a classical channel so that the input state can be reconstructed by the receiver using an appropriate unitary transform ( @xmath21 in fig . [ scheme ] ) @xcite . we now consider quantum teleportation of the vsp qubit using an entangled channel : @xmath22 , where @xmath23 _ etc_. the entangled channel at time @xmath6 is obtained using eq . as @xmath24 with which the teleportation is performed . the interaction time here should be half of the interaction time for the direct transmission because the each part of entangled channel travels half of the length for the direct transmission as depicted in fig . [ scheme ] . the bell - state measurement is performed to discriminate between the four bell states , @xmath25 and @xmath26 , as a joint measurement for the input state and the sender s part of the entangled channel . if the outcome of the bell - state measurement was @xmath27 , the required unitary transform is the @xmath28 operation that corresponds to @xmath29 phase shift . if the outcome was @xmath30 , the receiver does not need to do anything . however , a typical bell measurement scheme using linear optics and photodetectors @xcite can not discriminate the other two bell state , @xmath31 , so that the success probability is limited to 50% @xcite . the teleported state after an appropriate unitary transform is @xmath32 and the average success probability is @xmath33_{\rm avg}\\ & + { \rm tr } \left [ \langle\psi ^-_{\rm v}\| \left \ { { \|\psi_{\rm v } \rangle}{\langle \psi_{\rm v } \| } \otimes \rho^{ch}_{\rm v}(\tau/2 ) \right \ } \|\psi _ { \rm v}^-\rangle \right ] _ { \rm avg}=1/2 , \end{split}\ ] ] where average is taken over the bloch sphere of input state . interestingly , the success probability is not affected by photon losses even though the average fidelity is degraded as already implied in eq . . the average fidelity for the successful events obtained in the same way as the case of the direct transmission is @xmath34 where @xmath6 was replaced by @xmath35 for a comparison with the direct transmission . as shown in fig . [ single](a ) , the average fidelity of teleportation @xmath36 is always higher than that of the direct transmission @xmath37 . the figure also shows that @xmath36 goes below the classical limit 2/3 @xcite at @xmath38 while @xmath37 does so at @xmath39 . + a psp qubit is represented as @xmath40 , where @xmath41 and @xmath42 correspond to horizontally and vertically polarized states , respectively . using eq . , it is straightforward to find that a psp qubit in the direct transmission under photon losses evolves as @xmath43 and the average fidelity is obtained as @xmath44 . we then consider quantum teleportation for a psp qubit using an entangled channel : @xmath45 . the entangled channel at time @xmath6 obtained using eq . is @xmath46 where @xmath47 . here , the four bell states are @xmath48 and @xmath49 . here , the bell - state measurement can be performed using a 50:50 beam splitter , two polarizing beam splitters and four photodetectors @xcite . the teleported state after the bell - state measurement and a correct unitary transform is found to be identical to @xmath50 for the case of the direct transmission in eq . . again , only two of the bell states , @xmath51 , can be identified using linear optics @xcite and required unitary transforms are the identity operation and the @xmath28 operation that is realized with a half - wave plate . the average success probability is @xmath52_{\rm avg}\\ & + { \rm tr } \left [ \langle\psi ^-_{\rm p}\| \left \ { { \|\psi_{\rm p } \rangle}{\langle \psi_{\rm p } \| } \otimes \rho^{ch}_{\rm p}(\tau/2 ) \right \ } \|\psi _ { \rm p}^-\rangle \right ] _ { \rm avg}=t/2 . \end{split}\ ] ] in fact , the success probability in this case is identical for any input state , it is worth noting that the success probability of psp qubits is always lower than the success probability of vsp qubits . we obtain the average fidelity for the successful events as @xmath53 , where @xmath54 replaces @xmath6 for a comparison with the direct transmission as mentioned in sec . we plot @xmath55 and @xmath53 in fig . [ single](b ) . obviously , @xmath56 is always higher than @xmath57 , and @xmath56 goes below the classical limit 2/3 at @xmath58 while @xmath57 does so at @xmath59 . instead of single photons , superpositions of coherent states can be used for quantum information processing with their inherent advantages . the small - size implementations of superpositions of coherent states have been performed @xcite up to @xmath60 @xcite and arbitrary qubits were demonstrated @xcite . their large - size implementation is possible using the non - deterministic amplification scheme @xcite , a fock state with a large number @xcite , or multiple photon subtractions @xcite but it is yet experimentally challenging . coherent - state qubits not only lose their coherence but also undergo amplitude damping under photon losses . however , as the interaction time @xmath6 is the value known to the sender and the receiver , we can use @xmath61 as a dynamic qubit basis in order to reflect the amplitude damping as suggested in ref . adopting damped coherent states @xmath61 as the dynamic qubit basis , the time dependent target coherent - state qubit which we want the receiver to have is @xmath62 where @xmath63 , @xmath64 are some complex numbers and @xmath65 is normalization constant . to achieve this purpose , the sender actually transmit the state @xmath66 . it is straightforward to find that the initial state @xmath66 under direct transmission evolves to @xmath67 since the coherent states @xmath61 are not orthogonal to each other , we need an orthonormal basis which spans the input and the output states in order to obtain average fidelity on the bloch sphere . we take such a basis , @xmath68 , where the normalization factors are omitted . the input state is then represented as @xmath69 so that the average can be taken over @xmath70 and @xmath71 . the average fidelity between @xmath72 and @xmath73 is obtained as @xmath74 the average fidelity of teleportation was derived in ref . @xcite using the methods described in the previous subsections . to perform teleportation for a coherent - state qubit , an entangled coherent state @xmath75 is shared by the sender and the receiver . the bell - state measurement are supposed to discriminate between the states @xmath76 and @xmath77 where @xmath78 . this type of bell - state measurement can be performed using a 50:50 beam splitter and two photon - number - resolving detectors @xcite . the two measurement outcomes , @xmath79 and @xmath80 , require straightforward unitary transforms ( identity and @xmath29 phase shift ) and we take them as the successful events following ref . @xcite . by substituting @xmath6 in ref . @xcite with @xmath54 for a comparison with direct transmission as mentioned in secs . [ sec2 ] and [ sec3 ] , the average fidelity of teleportation for the successful events is @xmath81 where @xmath82 . the average success probability was shown to be @xmath83 @xcite . this is identical to that of vsp qubits , which is always higher than that of psp qubits . we plot @xmath84 and @xmath85 for several amplitudes of @xmath86 s in figs . [ diffcoh](a)-(c ) . if the amplitudes of coherent - state qubits are small as @xmath87 , @xmath88 is always higher than @xmath89 . however , as @xmath86 gets larger , the region where teleportation outperforms diminishes . the direct transmission outperforms for the weaker decoherence @xmath90 , whereas the teleportation is better for the stronger decoherence @xmath91 ( fig . [ diffcoh](d ) ) . we now compare vsp qubits , psp qubits , and coherent - state qubits under each information transfer scheme . the average fidelities for direct transmission and teleportation of vsp , psp , and coherent - state qubits with @xmath92 are plotted in figs . [ dirtel](a ) and [ dirtel](b ) . in both the schemes , vsp qubits are the most robust ones to decoherence in the region where comparing fidelities is meaningful , _ i.e. _ , above the classical bound @xmath93 . using direct transmission ( teleportation ) , the coherent - state qubits with small @xmath94 ( @xmath95 for the teleportation case ) outperforms psp qubits in the entire region of @xmath96 where the comparison is valid . however , as @xmath86 gets larger , the regions where coherent - state qubits outperforms psp qubits diminish . coherent - state qubits outperform psp qubits for the stronger decoherence @xmath91 , whereas psp qubits outperform coherent - state qubits for the weaker decoherence @xmath90 ( figs . [ dirtel](c ) and [ dirtel](d ) ) . considering the number of photons as a resource , we may compare psp qubits and coherent - state qubits when they have the same average photon number , _ i.e. _ , @xmath97 . the average photon number of input coherent - state qubits @xmath66 is @xmath98 where @xmath99 . therefore , the amplitude of coherent - state qubits for a comparison should be @xmath100 for @xmath97 to be the same to that of the psp qubits . the coherent - state qubits with the chosen amplitude @xmath101 always outperform psp qubits when the direct transmission is used as shown in fig . [ dirtel](a ) . however , when the teleportation protocol is used , psp qubits are more robust than coherent - state qubits with the chosen amplitude when decoherence is weak , and the opposite is true for strong decoherence ( fig . [ dirtel](b ) ) . the psp qubits in both the schemes eventually become the vacuum states , which leads their fidelities in fig . [ dirtel ] to vanish as @xmath102 . the coherent - state qubits and the vsp qubits become identical in the limit of @xmath103 as implied in figs . [ dirtel](a ) and [ dirtel](b ) . this is due to the fact that even and odd superpositions of coherent sates , @xmath104 ( without normalization ) , approach the vacuum and single photon , respectively @xcite . several different types of qubits have been suggested for optical quantum information processing and each of them has its own merits and limitations . a hybrid architecture using different types of qubits may be an efficient way to implement practical quantum information processing based on optical systems @xcite . in this context , it is important to make a thorough comparison among the different types of qubits in terms of transfer efficiencies in a lossy environment . we have compared three well - known different types of optical qubits , vsp , psp and coherent - state qubits , for information transfer via quantum teleportation and direction transmission under photon losses . of course , it should be noted that quantum teleportation always suffers lower success probabilities compared to the direct transmission if available resources are limited to linear optics elements and photon detectors in addition to the entangled pair @xcite . however , as far as fidelities are concerned , quantum teleportation always outperforms the direct transmission when vsp and psp qubits are used . the same applies to the coherent - state qubits when their amplitudes are as small as @xmath87 . on the other hand , the teleportation outperforms the direct transmission in the strong decoherence regions for the coherent - state qubits with large amplitudes . we have found that vsp qubits are the most robust ones against photon losses both for quantum teleportation and for the direct transmission . coherent - state qubits with small amplitudes ( @xmath94 for direct transmission and @xmath95 for teleportation ) are more robust to photon losses than psp qubits in optical quantum information transfer , whereas the coherent - state qubits with large amplitudes outperforms psp qubits only in the strong decoherence regions . this means that coherent - state qubits may be more effective than psp qubits for optical quantum information transfer particularly when photon loss effects are heavy . the success probabilities for teleportation of coherent - state qubits and vsp qubits ( _ i.e. _ , 1/2 regardless of losses ) are always greater than that of psp qubits ( _ i.e. _ , smaller than 1/2 under lossy effects ) . overall , vsp qubits are the most efficient for quantum information transfer under photon loss effects among the three types of qubits . in spite of our results clearly unfavorable to the psp qubits , the psp qubits may be preferred for certain applications such as quantum key distribution using single photons in which post - selection plays an important role @xcite . in this type of post - selection process , a result is simply discarded whenever any photon is missing at the final measurement . this is not so straightforward with the vsp or coherent - state qubits because the photon numbers of those qubits are inherently indefinite . in this paper , we have compared three types of optical qubits that can be represented by single - mode states . our results would provide useful and timely information for the development of practical optical quantum information processing . it would be an interesting future work to extend this comparison to optical qudits @xcite , continuous variable systems @xcite , and hybrid qubits @xcite . this work was supported by the national research foundation of korea ( nrf ) grant funded by the korea government ( msip ) ( no . 2010 - 0018295 ) and by the center for theoretical physics at seoul national university . 1 d. bouwmeester , j. pan , k. mattle , h. weinfurter , and a. zeilinger , nature ( london ) * 390 * , 575 ( 1997 ) . d. boschi , s. branca , f. de martini , l. hardy , and s. popescu , phys . lett . * 80 * , 1121 ( 1998 ) . e. knill , r. laflamme and g. j. milburn , nature * 409 * , 46 ( 2001 ) . p. kok , w. j. munro , k. nemoto , t. c. ralph , j. p. dowling , and g. j. milburn , rev . phys . * 79 * , 135 ( 2007 ) . lee and j. kim , phys . a * 63 * , 012305 ( 2000 ) . a. p. lund and t. c. ralph , phys . a * 66 * , 032307 ( 2002 ) . h. jeong and m. s. kim , phys . a * 65 * , 042305 ( 2002 ) . t. c. ralph , a. gilchrist , g. j. milburn , w. j. munro , and s. glancy , phys . a * 68 * , 042319 ( 2003 ) . s. j. van enk and o. hirota , phys . a * 64 * , 022313 ( 2001 ) . h. jeong , m. s. kim , and j. lee , phys . a * 64 * , 052308 ( 2001 ) . j. i. cirac , p. zoller , h. j. kimble , and h. mabuchi , phys . * 78 * , 3221 ( 1997 ) . p. van loock , laser & photonics review * 5 * , 167 ( 2011 ) . k. park , s .- w . lee , and h. jeong , phys . a * 86 * , 062301 ( 2012 ) . lee and h. jeong , phys . rev . a * 87 * , 022326 ( 2013 ) . o. morin , k. huang , j. liu , h. l. jeannic , c. fabre , and j. laurat , arxiv:1309.6191 . h. jeong , a. zavatta , m. kang , s .- w . lee , l. s. costanzo , s. grandi , t. c. ralph , and m. bellini , arxiv:1309.6192 . c. h. bennett , g. brassard , c. crepeau , r. jozsa , a. peres , and w. k. wootters , phys . lett . * 70 * , 1895 ( 1993 ) . m. takeoka , m. ban , and m. sasaki , j. opt . b : quantum semiclassical opt . * 4 * , 114 ( 2002 ) . s. l. braunstein , and h. j. kimble , phys . lett . * 80 * , 869 ( 1998 ) a. furusawa , j. l. srensen , s. l. braunstein , c. a. fuchs , h. j. kimble , and e. s. polzik , science * 282 * , 706 ( 1998 ) . k. park and h. jeong , phys . a * 82 * , 062325 ( 2010 ) . e. lombardi , f. sciarrino , s. popescu , and f. de martini , phys . * 88 * , 070402 ( 2002 ) . k. j. resch , j. s. lundeen , and a. m. steinberg , phys . . lett . * 88 * , 113601 ( 2002 ) . a. i. lvovsky and j. mlynek phys . . lett . * 88 * , 250401 ( 2002 ) . s. a. babichev , b. brezger , and a. i. lvovsky , phys . lett . * 92 * , 047903 ( 2004 ) . a. m. lance , h. jeong , n. b. grosse , t. symul , t. c. ralph , and p. k. lam , phys . a * 73 * , 041801(r ) ( 2006 ) ; h. jeong , a. m. lance , n. b. grosse , t. symul , p. k. lam , and t. c. ralph , phys . rev . a * 74 * , 033813 ( 2006 ) . h. jeong , w. son , m. s. kim , d. ahn , and . brukner , phys . a. * 67 * , 012106 ( 2003 ) . v. scarani , h. bechmann - pasquinucci , n. j. cerf , m. duek , n. ltkenhaus , and m. peev , rev . phys . * 81 * , 1301 ( 2009 ) . a. miranowicz , j. opt . b : quantum semiclassical opt . * 7 * , 142 - 150 ( 2005 ) . j. kim , j. lee , s .- w . ji , h. nha , p. m. anisimov , and j. p. dowling , arxiv:1012.5872 . s. lloyd and s. l. braunstein , phys . lett . * 82 * , 1784 ( 1999 ) .	we compare three different types of optical qubits for information transfer via quantum teleportation and direction transmission under photon losses . the three types of qubits are ( 1 ) qubits using the vacuum and the single - photon ( vsp ) states , ( 2 ) single - photon qubits using polarization degrees of freedom , _ i.e. _ , polarized single - photon ( psp ) qubits , and ( 3 ) coherent - state qubits that use two coherent states with opposite phases as the qubit basis . our analysis shows that the teleportation scheme outperforms the direct transmission for most of cases as far as fidelities are concerned . overall , vsp qubits are found to be the most efficient for both the direct transmission and teleportation under photon loss effects . the coherent - state qubits are more robust than psp qubits either when their amplitudes are small as @xmath0 or when photon loss effects are strong . our results would provide useful and timely information for the development of practical optical quantum information processing particularly in the context of hybrid architectures .
despite the intensive investigation of ruthenate - cuprates in the last ten years , a lot of questions about the nature of magnetic order and superconductivity , in particular their coexistence , are still unanswered @xcite . it is well established that in rusr@xmath2@xmath7cu@xmath2o@xmath3 ( ru1212@xmath7 , @xmath8 eu , gd , y ) compounds the magnetic ordering of ruthenium sublattice occurs at about 130 k. the ordering is predominantly antiferromagnetic ( afm ) with an easy axis perpendicular to the layers and with a weak ferromagnetic ( fm ) component parallel to the layers . while both , ruo@xmath2 and cuo@xmath2 layers may participate in the normal - state conductivity , the superconducting ( sc ) properties of ru1212@xmath7 compounds are associated only with the charge carriers in the cuo@xmath2 planes , and are , thus , strongly dependent on the concentration of these carriers . the related superconducting critical temperature @xmath5 ranges between 15 and 50 k , depending on the sample composition and/or sample preparation conditions . according to the thermopower and hall coefficient measurements , pure ru1212@xmath7 samples are expected to be intrinsically underdoped , with the effective hole concentration in the cuo@xmath2 planes nearly equal to @xmath9 @xcite . in high-@xmath5 superconductors , sc properties are strongly dependent on the effective hole concentration @xmath10 in the cuo@xmath2 planes . in the optimally doped systems , the effective hole concentration is estimated to be @xmath11 holes / cu . in this respect , in recent years there has been a lot of experimental effort to adjust the number of charge carriers in the cuo@xmath2 planes in ru1212@xmath7 samples using different substitutional impurities . the largest changes in @xmath5 are found when ru is partially replaced by cu ( @xmath12 k for 40% of ru replaced by cu @xcite ) . in several studies ruthenium was replaced by nb@xmath13,v@xmath14 sn@xmath15 , ti@xmath15 and rh@xmath16 @xcite . substitutions have also been made for other ions . for example , trivalent gadolinium was replaced by ca@xmath17 and ce@xmath15 @xcite or by isovalent y and dy @xcite . finally , divalent strontium was replaced by la@xmath16 and na@xmath18 @xcite . if the substitutional impurity lowers the total number of holes ( la@xmath16 for sr@xmath17 , ce@xmath15 for gd@xmath16 , nb@xmath13 for ru@xmath15 ) , @xmath5 is strongly depressed , indicating that @xmath10 is reduced well below @xmath19 . however , in the cases when valence counting of impurities would suggest a strong increase in @xmath10 ( na@xmath18 for sr@xmath17 , sn@xmath15 for ru@xmath13 ) , the increase in @xmath5 was not experimentally observed , contrary to the observation in common high-@xmath5 systems with substitutional impurities . this is rather similar to the observation in the underdoped yba@xmath2cu@xmath20o@xmath21 compounds where @xmath22 is nearly constant in a wide range of @xmath23 . the increase of @xmath5 was only observed in ca@xmath17 substitution for gd@xmath16 @xcite , but it can be attributed to the formation of ru@xmath0sr@xmath2gdcu@xmath24o@xmath3 phase . the magnetic ordering is , however , strongly influenced by the substitutions in all of the mentioned doping studies . there is still an open debate whether ruo@xmath2 layers are conducting or not . the magnetization @xcite and nqr @xcite results suggest that most of the electrons are localized resulting in ru@xmath15 ions while magnetoresistivity and hall measurements @xcite suggest the existence of conductivity in magnetically ordered ruo@xmath2 layers . we shall deal with this question in our doping study , where the number of charge carriers is modified by the replacement of ru ions with sn . in this paper a comprehensive microwave study on several ru1212eu and ru1212gd samples doped with sn is reported . one of the samples ( ru@xmath25sn@xmath26sr@xmath2gdcu@xmath2o@xmath3 ) is also characterized by transport , magnetoresistance and hall measurements , and the results are compared with our previous measurements @xcite . in this context , it is important to point out the controversy related to the role of the sn substitution on ru sites in ru1212gd samples . mclaughlin et al . @xcite and hassen and mandal @xcite reported , respectively , the increase and decrease of @xmath5 with doping . it is possible that this controversy reflects different ways to experimentally determine @xmath5 , or might be related to different methods of sample preparation . moreover , we shall revise the observation in the hall study ( fig . 2 in ref . @xcite ) in which the increase of the hall coefficient with increased sn content was reported . ru@xmath0sn@xmath1sr@xmath2@xmath7cu@xmath2o@xmath3 polycrystalline samples , where @xmath8eu or gd , were prepared by solid state synthesis , as described elsewhere @xcite . an elliptical @xmath27te@xmath28 copper cavity operating at 9.3 ghz was used for the microwave measurements . the sample was placed in the center of the cavity on a sapphire holder . at this position microwave electric field has its maximum . external dc magnetic field perpendicular to the microwave electric field was varied from zero up to 8 t. the temperature of the sample could be varied from liquid helium to room temperature . the measured quantity was @xmath29 , the total losses of the cavity loaded by the sample . it is simply related to the surface resistance of the material @xmath30 which comprises both , nonresonant resistance and resonant spin contributions . the details of the detection scheme are given elsewhere @xcite . resistivity , magnetoresistance and hall effect measurements were carried out in the standard six - contact configuration using the rotational sample holder and the conventional ac technique ( 22 hz , 1 ma ) , in magnetic fields up to 8 t. temperature sweeps for the resistivity measurements were performed with carbon - glass and platinum thermometers , while magnetic field dependent sweeps were done at constant temperatures which was controlled with a capacitance thermometer . the temperature dependence of the microwave surface resistance of various ru1212eu ( sn doping @xmath31 , 0.1 , 0.2 , 0.3 and 0.4 ) and ru1212gd ( @xmath31 , 0.2 ) compounds is shown in figs . [ rs_normalized](a ) and [ rs_normalized](b ) , respectively . the data are normalized at @xmath32 k , for comparison . we found that , for the samples of similar geometry , the absolute level of the normal - state absorption in the eu - based samples was systematically higher than the absorption in the gd - based samples . moreover , the surface resistance of the eu - based samples shows more pronounced rise as the temperature decreases towards @xmath33 ( @xmath5 , hereafter ) than it is the case in gd - based samples . the related crossover to the sc state occurs at lower temperatures . the sc crossover temperature region is rather broad in these data , resulting in a rather uncertain estimation of @xmath5 . this problem can be easily solved by taking into account that the superconductivity in the vicinity of @xmath5 is strongly dependent on magnetic fields . for this purpose , the difference between the microwave absorption in zero field and in @xmath34 t is plotted in fig . [ rs_differences ] . [ a considerably smaller signal of sn - doped sample in figs . [ rs_differences](b ) and [ rs_magnetic](d ) , with respect to the pure gd - based sample , is only due to the smaller geometry of the sample . ] both the magnetic and sc transitions can be detected here with a much better resolution than in fig . 1 . sn@xmath1sr@xmath2eucu@xmath2o@xmath3 and ( b ) ru@xmath0sn@xmath1sr@xmath2gdcu@xmath2o@xmath3 for various sn concentrations @xmath23 . ] t of ( a ) ru@xmath0sn@xmath1sr@xmath2eucu@xmath2o@xmath3 and ( b ) ru@xmath0sn@xmath1sr@xmath2gdcu@xmath2o@xmath3 for various @xmath23 . ] the magnetic ordering temperature @xmath4 of the ru lattice corresponds with small peaks clearly seen in fig . [ rs_differences ] . the dependence of @xmath4 on the sn content @xmath23 is shown in figs . [ phase](a ) and [ phase](b ) . @xmath4 is strongly suppressed with increasing @xmath23 in both the gd- and eu - based samples ( the magnetic critical temperature @xmath4 is reduced nearly by a factor of 2 for @xmath35 ) . this suppression seems to be a consequence of the reduced content of the ru magnetic ions ( dilution of the magnetic lattice ) and the increased disorder in the ru lattice . similar decrease in @xmath4 is found in the samples in which ru is replaced by nonmagnetic nb@xmath13 ions @xcite ( @xmath36 k for @xmath35 @xcite ) . ( open triangles ) and @xmath5 ( full circles ) on @xmath23 in ru@xmath0sn@xmath37sr@xmath38cu@xmath2o@xmath3 , for ( a ) @xmath8 eu and ( b ) @xmath8 gd . ] the sc ordering temperature @xmath5 is clearly seen in fig . [ rs_differences ] for all the samples , and its dependence on @xmath23 is further shown in fig . [ phase ] . the sudden drop in the difference of the microwave absorption in zero field and in @xmath34 t indicates the onset of the sc state in the grains . this effect can be associated with the intrinsic value of @xmath5 . on the other hand , the whole sample is expected to become superconducting at much lower temperatures where the 3d network of intergranular josephson junctions is established . it is clear that the influence of sn doping on @xmath5 is not nearly as dramatic as the influence on @xmath4 . in the gd - based samples , for example , the replacement of 20% of ru by sn does not appreciably change @xmath5 . in the eu - based samples , on the other hand , a slight reduction of @xmath5 with increasing @xmath23 is observed , but it is negligible in comparison with the related change in @xmath4 . eucu@xmath2o@xmath3 , ( b ) ru@xmath39sn@xmath40sr@xmath2eucu@xmath2o@xmath3 , ( c ) rusr@xmath2gdcu@xmath2o@xmath3 and ( d ) ru@xmath25sn@xmath26sr@xmath2gdcu@xmath2o@xmath3 measured at various temperatures . ] the sc state is further analyzed by measuring the magnetic field dependence of the microwave absorption . the results taken at different temperatures are shown in fig . [ rs_magnetic ] for two undoped samples and two sn - doped samples . at lower fields ( below 0.5 t ) , the microwave absorption is strongly field - dependent due to the intergranular superconductivity dominated by the 3d josephson network . at higher fields , all the josephson junctions are driven to the normal state , and superconductivity is localized only in the grains . the microwave absorption grows slowly in this field range as the number of vortices increases , and it is expected to reach the normal - state absorption at the upper critical field @xmath6 . the microwave magnetoresistance is therefore a useful tool for the estimation of upper critical fields . if one assumes that the depinning frequency is lower than the driving microwave frequency , it suffices to extrapolate the microwave absorption curves to the normal - state values and determine @xmath6 . however , this simple determination of @xmath6 is not possible in cases where the depinning frequency is comparable to the driving frequency . then , one has to take into account also the frequency shift . the details of the complete procedure are explaned in ref . the shape of the low - temperature microwave magnetoresistance curves of the gd - based samples [ @xmath41 k curves in figs . [ rs_magnetic](c ) and [ rs_magnetic](d ) ] points at the regime where the two frequencies are comparable to each other . at higher temperatures , on the other hand , the driving microwave frequency seems to be larger than the depinning frequency allowing the direct estimation of the upper critical fields . in the eu - based samples , the depinning frequency seems to be low enough even at low temperatures , so that the upper critical fields can be simply determined in the complete temperature region of figs . [ rs_magnetic](a ) and [ rs_magnetic](b ) . the upper critical fields @xmath6 estimated from the curves in fig . [ rs_magnetic ] ( and similar curves for other samples not shown here ) are plotted in fig . the comparison of @xmath6 estimated in two eu - based samples , @xmath42 and 0.3 , is shown in fig . [ bc2](a ) . the analogous comparison of @xmath6 in the gd - based samples with @xmath42 and 0.2 is given in fig . [ bc2](b ) . from these figures one can see that the upper critical fields in the gd - based samples are lower than in the eu - based samples , and , in both cases , the upper critical fields are higher in the doped samples than in the undoped samples . having in mind that higher @xmath6 means lower ginzburg - landau coherence length , which , in turn , depends on the electron mean free path , one concludes that the disorder in the undoped eu - based samples is higher than in the undoped gd - based samples , and that the sn doping introduces an additional disorder in both systems . the low - temperature ( @xmath41 and 10 k ) dependence of the upper critical field on the sn concentration @xmath23 in the eu - based samples is also shown [ fig . [ bc2](c ) ] . from fig . [ bc2](c ) , we can see that the upper critical field increases with @xmath23 for low to moderate concentrations , while for the highest concentration ( @xmath43 ) @xmath6 is suppressed . the drop of @xmath6 in the @xmath43 sample is related to the drop of @xmath5 in fig . [ phase](a ) . in order to clarify the role of the disorder introduced by sn doping , we extend the experimental study and analysis to transport coefficients . we have chosen the doped sample ru@xmath25sn@xmath26sr@xmath2gdcu@xmath2o@xmath3 which exhibits a sharp sc onset in fig . [ rs_differences](b ) , and measured its resistivity , magnetoresistivity and hall resistivity . sn@xmath26sr@xmath2gdcu@xmath2o@xmath3 measured in zero field ( black line ) and in @xmath34 t ( red line ) . inset shows the comparison of resistivity with the resistivity of the undoped sample ( thin line ) @xcite . ( b ) difference between the zero - field and 8 t resistivities . five temperature intervals labeled by ( a ) , ( b ) , ( c ) , ( d ) , and ( e ) correspond to subsets of mr curves in fig . [ mrall ] . ] figure [ rho](a ) shows the resistivity measured in zero magnetic field and in @xmath44 . in both cases , the resistivity at 200 k is 9.5 m@xmath45 cm , which is much lower than 15.5 m@xmath45 cm of the undoped sample @xcite . in zero field , one observes a relatively broad sc transition characterized by the change in the magnitude from 90% to 10% within the temperature interval @xmath46 k. such temperature dependence of the resisitivity in the vicinity of @xmath5 is usually attributed to the spontaneous vortex phase . the temperature interval characterizing the sc transition in the 8 t resistivity data is much broader than that of the zero - field data , and the zero resistance occurs only at 10 k , indicating that the pinning above 10 k is not strong enough to suppress the vortex motion , even for dc driving currents . the absence of strong pinning forces above 10 k justifies the estimation of @xmath6 from the magnetic field dependence of the microwave absorption given above [ fig . [ rs_magnetic](d ) ] . the onset of superconductivity in the grains could be seen better if the resistivity in @xmath47 is subtracted from the zero - field resistivity , as shown in fig . [ rho](b ) . the subtraction also reveals the existence of a broad maximum at the magnetic ordering temperature @xmath48 k. this peak can be further analyzed using magnetic field dependent resistivity measurements . sn@xmath26sr@xmath2gdcu@xmath2o@xmath3 in various temperature ranges . the values at 43.4 k and 40 k in ( e ) are divided by factors 3 and 10 , respectively . ] the relative transversal magnetoresistivity @xmath49/\rho(0,t)$ ] is shown in fig . [ mrall ] in a large temperature range from 140 k down to the superconducting transition . the curves are grouped in five subsets according to temperature intervals in fig . [ rho](b ) labeled by ( a ) , ( b ) , ( c ) , ( d ) , and ( e ) , exhibiting several qualitatively different physical situations . at temperatures well above @xmath4 [ fig . [ mrall](a ) ] , the magnetoresistivity is characterized by the negative quadratic - in - field behaviour similar to the magnetic - field dependence observed in dilute alloys containing uncorrelated magnetic impurities @xcite . with decresing temperature the correlations between magnetic ru ions start to play important role and the quadratic - in - field magnetoresistance transform into the linear - in - field behaviour . the same trend can also be seen for temperatures below @xmath4 [ fig . [ mrall](b ) ] , but not too close to @xmath5 . not surprisingly , the same linear - in - field behaviour was already observed in undoped and in la - doped samples in the vicinity of @xmath4 . just below @xmath4 , superposed to the negative linear magnetoresistivity , a small positive component develops at low fields , which is related to the afm ordering of the ru lattice . the same positive contribution is also observed in the longitudinal configuration ( i.e. @xmath50 ) . finally , it should be noticed that the negative magnetoresistivity of figs . [ mrall](a ) and [ mrall](b ) does not show any sign of saturation up to @xmath44 . figures [ mrall](c ) and [ mrall](d ) reveal a relatively complicated behaviour of the magnetoresistivity at temperatures between 57 and 47 k. the positive afm contribution decreases rapidly and vanishes below 53 k. below 49 k one observes positive contribution , presumably related to the sc fluctuations . the resulting magnetoresistivity , which is the sum of the positive sc and negative fm contribution , yields a local minimum . finally , below @xmath51 k , the positive contribution to the magnetoresistivity related to the sc ordering of the conducting layers starts to dominate . the sharp increase of the magnitude of the relative magnetoresistivity in this temperature region , shown in fig . [ mrall](e ) , reflects the sharp decrease of @xmath52 in fig . [ rho](a ) . sn@xmath26sr@xmath2gdcu@xmath2o@xmath3 at @xmath53 k ( circles ) and @xmath54 k ( squares ) . the hall resistivity of the undoped rusr@xmath2gdcu@xmath2o@xmath3 at @xmath55 k is shown by the dotted line for comparison @xcite . ( b ) temperature dependence of the low- ( open circles ) and high - field ( full squares ) hall coefficients of ru@xmath25sn@xmath26sr@xmath2gdcu@xmath2o@xmath3 . the averaged high - field hall coefficients of the pure rusr@xmath2gdcu@xmath2o@xmath3 and the la - doped rusr@xmath56la@xmath57gdcu@xmath2o@xmath3 are indicated by the arrows for comparison . ] so far we have established two seemingly opposed observations : the resisitivity of the sn - doped sample is lowered with respect to the pure compound , while the electron relaxation rates ( proportional to the resisitivity ) increases with the sn concentration @xmath23 ( as estimated from the @xmath23 dependence of @xmath6 ) . in order to resolve the problem , we have measured the hall resistance in which the relaxation rates cancell out in the first approximation , allowing in this way the estimation of the @xmath23 dependence of the effective number of charge carriers . the hall resistivity of ru@xmath25sn@xmath26sr@xmath2gdcu@xmath2o@xmath3 as a function of magnetic field is shown in fig . [ hall](a ) , for two temperatures below @xmath4 , and compared to the results of the undoped sample @xcite . a nonlinear increase of the hall resistivity , which is the combination of the ordinary and extraordinary hall contributions , and is a characteristic of magnetic metals , is observed at low fields . at high applied fields , a linear increase dominates , representing the ordinary hall contribution only . the magnitude of the extraordinary contribution is smaller than that in the undoped sample [ dotted line in fig . [ hall](a ) ] . the average values of the hall coefficient @xmath58 are further plotted in fig . [ hall](b ) for low and high fields . both , the low- and high - field slopes are smaller than for the pure compound . in particular , the absolute value of the ordinary @xmath59 is roughly 30% smaller than in the pure compound , indicating a substantial increase of the effective number of charge carriers with sn doping . before discussing these data in more detail ( sec . iv b ) , it is useful to recall the prediction for @xmath59 of the single - component free - hole model with the carrier concentration @xmath60 , which is found to explain well the low - temperature hall measurements in la@xmath61sr@xmath1cuo@xmath62 in the wide doping range @xmath63 @xcite . with @xmath64 @xmath65 being the primitive cell volume of ru1212gd , the result is @xmath66 m@xmath67c . the measured high - field hall coefficients indicated in fig . [ hall](b ) lead to @xmath68 , @xmath19 and @xmath69 for the pure , sn - doped and la - doped samples , respectively . in the following , we shall first show that the main changes in the zero - field resistivity , and in the ( high - field ) ordinary hall coefficent with sn doping ( which amounts to several tens of percent according to the inset of fig . [ rho](a ) and fig . [ hall ] ) , can be explained by using a single - component model for the conductivity tensor ( subsection [ ordinary ] ) . within this model , only cuo@xmath2 layers contribute to the observed conductivity , while the direct contributions of the ruo@xmath2 layers to the total conductivity are small and can be safely neglected . then , we shall argue , in subsection [ magnetic ] , that this single - component model fails to account for the observed contribution to the magnetoresistivity ( changes of the order of a few percent in fig . [ rho ] ) , and the ( low - field ) hall coefficient . in order to successfully explain these experimental features , one has to go beyond the single - component model and to take into account a presumably small number of itinerant electrons in the ruo@xmath2 layers . the motivation for such two - step analysis of our data comes from the results of the zero - field nmr ( nqr ) experiments in the pure ru1212gd @xcite and the related band structure calculations @xcite . the number of electrons in ruo@xmath2 layers is equal to the number of holes doped into the cuo@xmath2 plane . naively , ru@xmath15 ion can be visualized as a composition of ru@xmath13 ion and one extra electron . in this picture , the basic question is as to which extent this extra electron is localized to the parent ru@xmath13 ion . in ru - nqr experiments , two well - distinguished signals are found , corresponding to ru@xmath13 with the spin state @xmath70 and ru@xmath15 with the spin state @xmath71 . these results suggest that most of the electrons are localized ( leading to ru@xmath15 with the spin state @xmath72 ) , which means that the number of itinerant ( delocalized ) electrons in the ruo@xmath2 layers is small . in nqr , it is also found that the magnetic order in the ruo@xmath2 layers coexists with the sc order in the cuo@xmath2 planes down to @xmath73 k. finally , the nqr frequency @xmath74 measured on @xmath75cu nuclei @xcite , which is close to @xmath76 measured in the underdoped yba@xmath2cu@xmath20o@xmath77 compounds @xcite , puts ru1212@xmath7 systems into the high-@xmath5 superconductors with an intermediate copper - oxygen hybridization and , consequently , an intermediate value of @xmath5 @xcite . the band structure calculations @xcite , on the other hand , suggest the itinerant character of electrons in the ruo@xmath2 layers , rather than the localized character , but all other conclusions of ref . @xcite are consistent with the nqr observations . namely , there are three bands crossing @xmath78 , the first two are related to two cuo@xmath2 layers , while the third one is related to the ruo@xmath2 layer . the magnetic ordering and the calculated magnitude of the magnetic moments of ru ions are also consistent with the local magnetic fields measured by nqr @xcite . in the present analysis , the splitting between two hole - like cuo@xmath2 bands is neglected and the afm fluctuations in the cuo@xmath2 planes are taken into account @xcite . this gives the effective concentration of charge carriers ( holes ) in the cuo@xmath2 layers @xmath79 to be proportional to @xmath10 ( @xmath80 ) rather than to @xmath81 , and with the band mass close to the free electron mass . thus , @xmath79 nearly measures the hole concentration in the cuo@xmath2 bands with respect to the half - filling . the effective concentration of itinerant electrons in the ruo@xmath2 layers @xmath82 is assumed to be small in comparison with the value @xmath83 characterizing the completely delocalized case . in the presence of the magnetic background of ru@xmath13/ru@xmath15 ions and external magnetic fields the conductivity tensor in question is given by @xcite @xmath84 , \nonumber \\ \sigma_{xy } & = & -\sigma_{yx } = \frac{e^2}{m}\omega _ 0 \big[\tau _ { \rm cu}^2 \tilde{n}^{\rm cu}_{xy } + \tau _ { \rm ru}^2 \tilde{n}^{\rm ru}_{xy } \big ] + \sigma_{xy}^{\rm ex}. \label{eq1}\end{aligned}\ ] ] the first term in the off - diagonal component @xmath85 comes from the lorentz force and is proportional to @xmath86 ( @xmath87 is the cyclotron frequency ) . the second term , @xmath88 , includes all extraordinary contributions and is proportional to the magnetization of the ru lattice , which means that the dependence on @xmath86 is given in terms of the brillouin function @xmath89 . according to ref . @xcite , the contributions of two bands in eqs . ( [ eq1 ] ) can be regarded as decoupled ( i.e. the hybridization between the cuo@xmath2 and ruo@xmath2 bands is negligible ) . the structure of the effective numbers @xmath90 , @xmath91 cu , ru and @xmath92 , for this case is given in the appendix . in the absence of magnetic fields , for example , we obtain @xmath93 and @xmath94 , with @xmath79 and @xmath95 defined above . @xmath96 and @xmath97 are , respectively , the relaxation times of holes and electrons averaged over two spin projections ( see appendix ) . in the absence of magnetic fields , it is reasonable to assume that not only @xmath98 is small , when compared to @xmath99 , but also the relaxation rate in the ruo@xmath2 layers @xmath100 is much larger than @xmath101 , due to the presence of magnetic ions and the substitutional impurities in the ruo@xmath2 layers . thus , @xmath102 for @xmath103 . but , as mentioned above , the main effects in the magnetoresistivity are expected to come from the field dependence of @xmath104 [ labeled hereafter by @xmath105 . for weak magnetic fields , the general expressions for the hall resistivity and magnetoresistivity @xmath106 lead to @xmath107 i.e. @xmath108 + ( m / e^2)\sigma_{xy}^{\rm ex ; ru } } { \big(\tau_{\rm cu } n_{\rm cu } + \tau_{\rm ru } n_{\rm ru } \big)^2 } , \nonumber \\ \rho_{xx } & \approx & \frac{m}{e^2 } \frac{1}{\tau_{\rm cu } n_{\rm cu } + \tau_{\rm ru } n_{\rm ru}}. \label{eq4}\end{aligned}\ ] ] the hall coefficient is given by @xmath58 . the anomalous contributions ( the negative magnetoresistivity and the extraordinary hall coefficent ) are hidden here in @xmath109 and @xmath110 , and we put @xmath111 in the rest of this subsection . in the sn doped samples , the replacement of any ru atom by sn introduces one extra hole in the system , which is redistributed between ruo@xmath2 and cuo@xmath2 layers . the part of the extra charge which remains in the ruo@xmath2 layers decreases the effective number of itinerant electrons @xmath112 ( @xmath113 ) , thus increasing both the resistivity and hall resistivity . one may say that it gives rise to the conversion of ru@xmath15 into ru@xmath13 . the part of the extra charge transferred into the cuo@xmath2 layers increases the effective number @xmath79 ( @xmath114 . the experimentally determined decrease of the normal - state resistivity and the ( high - field ) hall coefficent with the sn doping ( figs . [ rho ] and [ hall ] ) shows that a substantial portion of the doped holes are indeed transferred to the cuo@xmath2 layers and that the small changes in already small @xmath112 can be neglected . this scenario is expected to hold for not too large number of sn impurities ( @xmath115 ) . our results should be contrasted to the sn doping study of mccrone et . al ( fig . 2 in ref . @xcite ) , where the increase of resistivity and the hall coefficient was observed for @xmath116 . as mentioned above , the correlation between @xmath79 and the superconducting critical temperature @xmath5 is expected to be similar to the correlations found in the isostructural yba@xmath2cu@xmath20o@xmath21 compounds . the main difference is that a very complicated role of the cuo chains in supporting superconductivity in the cuo@xmath2 layers is played here by the magnetically ordered ruo@xmath2 layers . indeed , the dependence of @xmath5 on the sn doping @xmath23 ( shown in fig . 3 ) is similar to the observation in the underdoped yba@xmath2cu@xmath20o@xmath21 compounds ( @xmath5 is nearly constant in a wide range of @xmath23 , @xmath117 k in yba@xmath2cu@xmath20o@xmath21 ) . similar scenario holds also for other ru1212@xmath7 systems . for example , the replacement of sr@xmath17 by la@xmath16 increases the total number of conduction electrons in the system . a substantial part of the doped electrons are transferred to the cuo@xmath2 layers , reducing the effective number @xmath79 below the critical value required for superconductivity , in agreement with experiments . in rusr@xmath56la@xmath57gdcu@xmath2o@xmath3 , the high temperature resistivity and the high - field hall resistivity are found to be increased by roughly 60 - 70% with respect to the pure compound , and the superconductivity is completely suppressed @xcite . the same effect of reducing the effective number @xmath79 is obtained by the partial replacement of ru ions by nb @xcite . the magnetoresistivity data of fig . [ mrall ] give us the indirect way to study the magnetic properties of the ru lattice . in the model described above , we can write @xmath118 with @xmath119 and @xmath120 being , respectively , the zero - field total conductivity and the zero - field conductivity of the ruo@xmath2 subsystem . the first factor is nearly proportional to @xmath121 , and represents an enhancement factor dependent on sn doping . it grows with decreasing @xmath79 , as can be easily seen from our data taken in three compounds : ru@xmath25sn@xmath26sr@xmath2gdcu@xmath2o@xmath3 , ru@xmath122sr@xmath2gdcu@xmath2o@xmath3 , and ru@xmath123la@xmath57sr@xmath2gdcu@xmath2o@xmath3 . the second factor in eq . ( [ eq5 ] ) , @xmath124 , is the magnetoresistivity of the isolated ruo@xmath2 subsystem , and it is expected to be similar to that of magnetic metals @xcite . precisely , the negative quadratic - in - field magnetoresistivity observed at temperatures well above @xmath4 has to be contrasted to the magnetoresistivity of common high-@xmath5 superconductors which is very small and positive @xcite . this high temperature behaviour can be understood as a clear evidence of the exchange interaction between itinerant electrons in the ruo@xmath2 layers and the localized ru spins @xcite . the afm interactions between localized spins become important at low enough temperatures leading to the linear - in - field behaviour with maximal slope at @xmath125 [ figs . [ mrall](a)-(c ) ] . we recall that exactly the same temperature evolution of the magnetoresistivity was found in ref . @xcite for small concentration of conduction electrons and low stability of the fm state of the localized spins . hence , we believe that expression ( [ eq5 ] ) includes all relevant processes in the magnetoresistivity of ruthenate cupartes in the nonsuperconducting temperature range . taking into account that @xmath126 , one can rewrite eq . ( [ eq5 ] ) in the form : @xmath127 for the pure ( rusr@xmath2gdcu@xmath2o@xmath3 ) , sn - doped ( ru@xmath25sn@xmath26sr@xmath2gdcu@xmath2o@xmath3 ) and la - doped ( rusr@xmath56la@xmath57gdcu@xmath2o@xmath3 ) samples at their respective magnetic ordering temperatures . ] where @xmath128 . figure [ colossal ] shows the field dependence of the quantity @xmath129 for the three aforementioned samples at their respective magnetic ordering temperatures . if one assumes , based on the previous subsection , that @xmath130 and @xmath131 , the magnetoconductivity of ruthenium layers is found to exceed 18% at @xmath44 . the magnitude of this effect is comparable to the colossal magnetoresistivity of the manganite perovskites . in conclusion , a microwave study of sn doping in ru1212 samples has shown that magnetic ordering is strongly influenced by introduction of tin into ruo@xmath2 layers , while the superconducting critical temperature remains practically unchanged . the magnetic field dependence of the microwave absorption has shown that the upper critical field in doped samples is increased , probably due to the increased disorder . however , in spite of increased disorder , the dc resistivity measurement in ru@xmath25sn@xmath26sr@xmath2gdcu@xmath2o@xmath3 is lower than in the respective pure compound . we conclude that it is due to the increased number of holes in the cuo@xmath2 layers . this conclusion is also confirmed by hall resistivity measurements presented in this paper . in ru@xmath25sn@xmath26sr@xmath2gdcu@xmath2o@xmath3 magnetoresistivity develops from quadratic - in - field behavior ( characteristic for uncorrelated magnetic impurities ) at high temperatures to linear - in - field behavior close to @xmath132 , indicating significant correlations between magnetic ru ions . below @xmath4 , a small positive component of magnetoresistivity , related to the afm ordering of the ruo@xmath2 lattice , is superposed to the negative linear magnetoresistivity . as the temperature is lowered below @xmath133 , the field - induced ferromagnetic order seems to prevail . at the onset of superconductivity ( @xmath134 ) both , positive superconducting and negative ferromagnetic contributions to the same magnetoresistivity curve occur . we have shown that the single component model for the conductivity tensor suffices to explain the variation of the normal - state resistivity and the high - field hall coefficient with doping . the main conductivity channel is through cuo@xmath2 layers and the various dopings simply change the number of conducting holes in these layers . however , the low - field hall resistivity and the magnetoresistivity provide evidence that there is a small conduction in magnetically ordered ruo@xmath2 layers . quantitative analysis of the observed negative magnetoresistance , leads to the conclusion that small number of delocalized electrons display a colossal magnetoresistance which exceeds 18% . this is a remarkable result . we acknowledge funding support from the croatian ministry of science , education and sports ( projects no . 119 - 1191458 - 1022 `` microwave investigations of new materials '' , no . 119 - 1191458 - 0512 `` low - dimensional strongly correlated conducting systems '' and no . 119 - 1191458 - 1023 `` systems with spatial and dimensional constraints : correlations and spin effects '' ) , the new zealand marsden fund , the new zealand foundation for research science and technology , and the alexander von humboldt foundation . in the multiband models with the interband hybridization negligible ( ru1212@xmath7 systems are the example ) , the elements of the conductivity tensor are given by the sum of the intraband terms @xmath135 ( @xmath91 cu , ru in the present case ) . the presence of magnetic background and external magnetic fields makes , in the first place , that the relaxation rates of conduction electrons @xmath136 depend on spin , and also introduces the extraordinary contributions to the off - diagonal elements of the conductivity tensor @xmath137 , as discussed in the main text . in the first approximation ( the case of noninteracting magnetic impurities @xcite ) we can write @xmath138 , where @xmath139 is a constant and @xmath140 is the total magnetization of the ruo@xmath2 layer ) . thus , in the ru1212@xmath7 systems , at temperatures well above @xmath4 , we have @xmath141 and @xmath142 , with @xmath143 and @xmath144 given by @xcite @xmath145 ^ 2 \tau_{i \sigma } ( { \bf k } ) } { 1 + \tau_{i \sigma}^2 ( { \bf k } ) \omega_{i0}^2 ( { \bf k } ) } ( -)\frac{\partial f_{i \sigma } ( { \bf k})}{\partial \varepsilon_{i \sigma } ( { \bf k } ) } , \nonumber \\ \tau_i^2 \tilde{n}_{xy}^i & \approx & \frac{1}{v } \sum_{{\bf k } \sigma } \frac { m v_x^i ( { \bf k})\big[v_x^i ( { \bf k } ) \gamma_{yy}^i ( { \bf k } ) - v_y^i ( { \bf k})\gamma_{xy}^i ( { \bf k } ) \big ] } { 1 + \tau_{i \sigma}^2 ( { \bf k } ) \omega_{i0}^2 ( { \bf k } ) } \nonumber \\ & & \times \tau_{i \sigma}^2 ( { \bf k})(-)\frac{\partial f_{i \sigma } ( { \bf k})}{\partial \varepsilon_{i \sigma } ( { \bf k})}. \label{eqa1}\end{aligned}\ ] ] @xmath146 is the electron group velocity , @xmath147 , and @xmath148 are the elements of the inverse effective - mass tensor [ @xmath149 . if magnetic fields are weak and the exchange interaction between the conduction holes and localized ru spins is negligible , the field dependence in the denominators in ( [ eqa1 ] ) can be neglected , and the sum over spin projections gives for the averaged relaxation times of itinerant electrons the expression @xmath150 $ ] , resulting in @xmath151 @xcite . according to eq . ( [ eq5 ] ) , this also leads to the total magnetoresistivity which is negative and quadratic - in - field . the effects of interactions between localized spins on the conductivity tensor at different temperatures are expected to be similar to that found in ref .	ru@xmath0sn@xmath1sr@xmath2eucu@xmath2o@xmath3 and ru@xmath0sn@xmath1sr@xmath2gdcu@xmath2o@xmath3 have been comprehensively studied by microwave and dc resistivity and magnetoresistivity and by the dc hall measurements . the magnetic ordering temperature @xmath4 is considerably reduced with increasing sn content . however , doping with sn leads to only slight reduction of the superconducting critical temperature @xmath5 accompanied with the increase of the upper critical field @xmath6 , indicating an increased disorder in the system and a reduced scattering length of the conducting holes in cuo@xmath2 layers . in spite of the increased scattering rate , the normal state resistivity and the hall resistivity are reduced with respect to the pure compound , due to the increased number of itinerant holes in cuo@xmath2 layers , which represent the main conductivity channel . most of the electrons in ruo@xmath2 layers are presumably localized , but the observed negative magnetoresistance and the extraordinary hall effect lead to the conclusion that there exists a small number of itinerant electrons in ruo@xmath2 layers that exhibit colossal magnetoresistance .
cells display complex nonlinear and time - scale dependent rheological properties @xcite . a broad range of relaxation timescales results in power - law spectra for the frequency dependence of the linear viscoelastic response @xcite . under nonlinear loading conditions , cells can display apparently contradicting behaviors , ranging from fluidization to reinforcement @xcite . understanding these mesoscale behaviors in terms of underlying non - equilibrium processes , such as cytoskeletal remodeling , motor activity or reversible crosslink binding or folding , remains an important theme in current biomechanical research . over recent years , reconstituted f - actin networks have become a popular model system in which these phenomena can be studied in detail . much of previous research in this field focused on the frequency - dependent rheology of _ permanently crosslinked _ filament networks @xcite . key questions revolve around the high - frequency modulus and its dependence on frequency @xmath0 @xcite , the nature of network deformations ( affine vs. non - affine ) at intermediate freuqencies @xcite and the nonlinear response properties of the network @xcite . theoretical models @xcite and simplified simulation schemes @xcite have been proposed that aim at explaining one or the other of these non - trivial features . in these studies , the filaments and their mechanical and thermal properties are assumed to dominate the effective rheology of the system . this may be different in f - actin networks crosslinked with the rather compliant crosslinking protein filamin . experimental and theoretical work @xcite suggest a second , crosslink - dominated regime , where network rheology is set by the crosslink stiffness , while filaments effectively behave as rigid , undeformable rods . in between these two extreme scenarios a proper treatment would have to consider the full interplay between crosslink and filament mechanical properties . this has not been addressed theoretically before . recent experiments have indicated that at low driving frequencies effects due to crosslink binding become important . this is evidenced , for example , as a peak in the loss modulus @xmath1 @xcite or a broad distribution of time - scales leading to an anomalous scaling with frequency , @xmath2 @xcite . theoretical modelling in this field are only beginning to emerge @xcite . some of the pertinent problems are : what is the force on a crosslink and how does it depend on network deformation ? how does this affect binding ? once unbound , how does this affect the network ? a noteworthy recent development is the phenomenological `` glassy wormlike chain '' model @xcite . in that approach , filament - filament interactions , as for example mediated by specific crosslinking , are not modeled explicitly , but are assumed to lead to an exponential stretching of the single filament relaxation times . network deformation is accounted for by a pre - stretching tensile force in the filament . the same force is assumed to enter the unbinding rate constant of the crosslinks via a bell - like model @xcite . here , we go beyond previous studies by investigating the interplay between filament and crosslink elasticity and its effects on the crosslink binding behavior . we will present a simple model that accounts for the individual crosslinks , their mechanical properties as well as their binding state . the model is a fully thermodynamic treatment , where crosslink binding is equilibrated for a given network deformation . no rate effects will be considered . as a result of our analysis , we will be able to selfconsistently calculate the nonlinear elastic modulus of the network , which incorporates as key ingredient the ensuing tendency for crosslinks to unbind under load . the manuscript is structured as follows . in section [ sec : model ] we will define our model and relate it to existing approaches in the literature . in the model the filament network is represented as an effective elastic medium with a given , fixed modulus @xmath3 . we will introduce a hamiltonian that describes the properties of a test filament embedded in this medium . in section [ sec : results ] we will present results of metropolis monte - carlo simulations for the hamiltonian introduced . in section [ sec : theor - fram ] a theoretical framework will be developed that allows some analytical results to be obtained . finally , in section [ sec : selfc - determ - medi ] we will discuss the question of how to obtain in a self - consistent way the stiffness of the effective medium in terms of the response properties of the test filament . we will consider the properties of a test filament crosslinked into a network . the filament is described in terms of the worm - like chain model . in `` weakly - bending '' approximation the bending energy of the filament can be written as @xmath4 where @xmath5 is the filament bending stiffness and @xmath6 is the transverse deflection of the filament from its ( straight ) reference configuration at @xmath7 . in these expressions @xmath8 is the arclength , @xmath9 $ ] , and @xmath10 is the length of the filament . the effect of the surrounding network is to confine the test filament to a tube - like region in space . in this way the actual network is substituted by an effective potential that acts on the test filament . a convenient potential is the harmonic tube @xmath11 where @xmath12 is the strength of the confinement and @xmath13 is the tube center , which may or may not be different from the reference configuration of the filament . the tube potential is a convenient representation of many - body network effects on the test filament . however , the parameters entering the potential ( here @xmath14 and @xmath15 ) and their relation to network deformation and mechanical properties are usually unknown . ideally , these parameters can be determined in a self - consistent way from the analysis of the test filament . in the following , we will present a theoretical framework , where such a self - consistent determination is partly possible ( see fig . [ illustration ] for a short description ) . in this model we assume the network to be represented by an effective medium that couples to the test filament only at the crosslinking points , @xmath16 where @xmath17 is the crosslink stiffness . @xmath18 is the total number of crosslinking sites . being interested in the effects of reversible crosslink binding , we also include an occupation variable @xmath19 . if the crosslink is bound at site @xmath20 , we have @xmath21 , if it is unbound then @xmath22 . below , we will also introduce ( and calculate self - consistently ) the modulus @xmath3 of the effective medium . the crosslink stiffness @xmath17 in equation ( [ eq : kx ] ) then has to be substituted by an effective stiffness @xmath23 , which contains contributions from both , the crosslinks and the effective medium . for the time being we set @xmath24 and thus @xmath25 . as we want to discuss the role of network deformation on the test filament , we need to know how a macroscopically applied strain field couples to the single filaments . if the effective medium can be thought of as strictly homogeneously elastic then the local strain felt by the filaments would be identical to the macroscopic strain @xmath26 imposed by the experimental device . with this assumption of `` affine '' deformations any network deformation couples to the end - to - end distance of the filament and leads to its extension or compression . rheological properties are then governed by the resistance of filaments to changes in their end - to - end distance@xcite . an alternative point of view is based on the fact , that filaments are anisotropic elastic objects @xcite , where bending is a much softer deformation mode than stretching , and a stretching stiffness @xmath27 we find @xmath28 as the persistence length @xmath29 is usually much larger than the relevant filament length @xmath30 . ] . this , together with local structural heterogeneity , may in fact lead to a highly nonaffine response where network properties are determined by the resistance of filaments to bending deformations @xcite . it has been pointed out in @xcite that the filament length @xmath10 may play the role of a lower cut - off at which the affine assumption breaks down . this breakdown is due to inextensibility and geometric correlations that develop along essentially straight filaments . the actual deformations on the smaller scale of the crosslink ( or the mesh - size ) then follow from the local network structure . adopting this latter point of view we assume that macroscopic and homogeneous network deformations ( as measured by a strain @xmath26 ) locally lead to an inhomogeneous distortion of the effective medium that can be accounted for by a strain - dependent tube center line @xmath31 . the occurence of the filament length @xmath10 in this equation is a direct consequence of its role as a nonaffinity length - scale . for our purposes the scaling with @xmath10 is not decisive , however . what is more important here is that nonaffine deformations of the tube increase with network strain @xmath26 and are slaved to the local network structure . in terms of our discrete crosslinking points this means that one head of the crosslink follows the effective medium deformation at @xmath32 , while the second head remains on the filament at @xmath33 . whenever the filament does not follow the medium , it is the crosslinks that have to deform by @xmath34 leading to an elastic energy @xmath35 . the tube potential is therefore identified with the crosslink deformation energy . the model thus describes an interplay between crosslink deformation and filament bending . a macroscopic strain applied to the background medium leads to a frustration effect between crosslink and filament deformations . with the possibility of crosslink unbinding , this frustration is avoided . the price to pay is the binding enthalphy @xmath36 , which is assumed to favour the bound crosslink state with @xmath21 . a similar competition between filament and crosslink energies is present in ordered arrays of parallel filaments or filament bundles @xcite . in that case , the role of the medium strain is taken by a deformation of the bundle as a whole . this deformation induces filament sliding and leads to a frustration effect between filament stretch and crosslink shear . subsequently , and to relieve this frustration , a discontinuous crosslink unbinding transition is observed . interestingly , this has the form of an unzipping transition @xcite . crosslinks unbind first at places where bundle deformation is largest . an interface then moves rapidly to the other end of the bundle . in the following we will analyze the model defined in equations ( [ eq : hb ] ) to ( [ eq : kx ] ) with the aim of understanding the combined effects of filament deformation and crosslink binding in response to a network strain . we will discuss some thermodynamic parameters as a function of strain , in particular the average crosslink occupation , @xmath37 and the filament energy @xmath38 . the latter will provide the connection to the stiffness of the network and can be determined in rheological experiments . for the monte - carlo simulations we represent the filament by a one - dimensional lattice . to each lattice site @xmath39 the pair @xmath40 is attached , denoting the local transverse displacement @xmath33 and the crosslink occupation variables @xmath41 . as monte carlo moves we use single site displacements and crosslink binding / unbinding moves , the latter attempted with ten percent probability . in a displacement move , a lattice site @xmath20 is selected randomly , and the current displacement @xmath33 of that site is replaced by @xmath42 , with @xmath43 drawn uniformly randomly , and @xmath44 is the discretization length , which serves as unit of length here . the new displacement @xmath45 is accepted with probability @xmath46 $ ] , where @xmath47 is the energy difference between initial and final state . during a crosslink move , a bond is selected randomly , and the corresponding occupation variable @xmath41 is flipped ( @xmath48 ) . the new state is accepted with probability @xmath49 $ ] , with @xmath50 the crosslink chemical potential , @xmath51 the change in the number of crosslinks . in the following we show data where the filament bending stiffness @xmath52 serves as unit of energy . the chemical potential is @xmath53 . the inverse temperature is @xmath54 such that the persistence length is @xmath55 , measured in units of lattice sites @xmath56 . the filament length is taken to be @xmath57 , i.e. @xmath58 . for simplicity , we assume the filament to have zero deflection at its ends , @xmath59 . the tube potential is taken as @xmath60 with @xmath61 , which corresponds to the longest possible wavelength compatible with the chosen boundary conditions . in fig.[n.energy ] we monitor the average crosslink occupation @xmath62 as well as an average energy @xmath38 , which is obtained by minimizing the total elastic energy for given crosslink occupation . we vary the network strain @xmath26 as well as the crosslink stiffness @xmath17 . it can readily be seen that crosslink stiffness has a dramatic effect on the thermodynamic state of the system . for mechanically weak crosslinks network strain @xmath26 has no strong influence on the binding state . only few crosslinks unbind upon increasing network deformation . in this regime the crosslinks are not strong enough to enforce filament bending . the elastic energy is small , and primarily stored in the crosslinks . as a consequence , the energy of the bound state is raised and the statistical weight is shifted towards the unbound state . this is the regime discussed in @xcite in the context of filamin - crosslinked f - actin networks . note , that in that model no crosslink unbinding is accounted for . the crosslinks are modeled as nonlinear elastic elements , leading to significant stiffening of the network under strain . we have checked that incorporating nonlinear crosslink compliance in our model reproduces this behavior . moreover , with the possibility of crosslink unbinding , this stiffening due to crosslink mechanics generally competes with a softening effect due to crosslink unbinding . when the crosslinks are sufficiently strong , their deformation energy starts to compete with the filament bending energy . now crosslinks can force filaments into deformation and the elastic energy is mainly stored in the filaments . at large network strains this energy is too high , however , and unbinding becomes favourable . this is evident as a discontinuous unbinding transition , where nearly all remaining crosslinks unbind simultaneously . associated with such a transition is a free - energy barrier . the escape time over this barrier sets the time - scale of relaxation of the imposed deformation mode . this will be exponential in the number of crosslinking sites @xmath18 , reminiscent of what has been proposed in @xcite . theoretical progress can be made in the continuum limit of equations ( [ eq : hb ] ) to ( [ eq : kx ] ) . making the mean - field assumption @xmath63 , the energy of the test filament can be written as @xmath64 going to fourier space , @xmath65 , the @xmath66 can be integrated out . the resulting effective free energy @xmath67 can be written as @xmath68 with a deformation - dependent part @xmath69 @xmath70 a deformation amplitude @xmath71 and the scale for crosslink density @xmath72 . the prefactor @xmath73 is an entropic contribution that specifies the entropic cost of binding to crosslinks . clearly , binding suppresses bending undulations and therefore reduces the entropy stored in these modes . the interplay between entropy reduction and @xmath50-dependent enthalpy gain may lead to a thermal unbinding transition @xcite . for the case considered here such an unbinding transition is not relevant . instead , we want to focus on the deformation - dependent part @xmath69 . just as in the simulations , we make the single - mode assumption , @xmath74 and @xmath75 . then it is easy to see that the free energy has two saddle - points at @xmath76 and @xmath77 and a discontinuous transition between them at @xmath78 . this condition gives a critical network strain of @xmath79 , which compares well with the actual transition as seen in figure [ n.energy ] . this analysis predicts a discontinuous unbinding transition for any value of crosslink stiffness @xmath17 , in apparent disagreement with the simulations . it turns out that this is a result of the saddle - point approximation . in a more refined treatment , which also includes an `` entropy of mixing '' term , @xmath80 , the discontinuous transition is destroyed when the crosslinks are sufficiently soft @xcite . finally , let us comment on the use of the single - mode assumption for the tube center @xmath81 . the precise form of @xmath82 and its dependence on network strain @xmath26 is , in principle , unkown and one would like to calculate it selfconsistently . as this is unfeasible , one has to rely on assumptions and physical plausibility . conceptually , the tube deformation in response to network strain @xmath26 represents the missing link between affine deformations on scales larger than the filament length and the actual crosslink motion on the scale of the mesh - size . as such it will be sensitive to the local structure of the network . in @xcite it has been shown how such a link can be constructed in terms of local binding angles and the mesh - size distribution . for our purposes we note , that using one or few higher modes will not fundamentally alter the proposed picture of continuous vs. discontinuous crosslink unbinding . the necessary condition for such a feature to persist is a free energy contribution @xmath69 that grows slower than linear with crosslink occupation @xmath83 . in this case the linear contribution @xmath84 from the binding enthalpy will eventually take over . with one or a finite number of modes present the free energy will eventually saturate @xmath85 . this happens , when the filament is strongly constrained by the crosslinks to precisely follow the tube centerline @xmath86 . clearly , binding even more crosslinks can not lead to a more efficient confinement . a similar result is obtained , if one assumes infinitely many modes , with an amplitude that depends on mode - number as @xmath87 , i.e. like a thermalized bending mode of a worm - like chain . in this case , the free energy does not saturate but asymptotically grows like @xmath88 , i.e. also slower than linear . in the previous sections we assumed the test filament to be coupled to a medium of infinite stiffness . in reality the medium itself is made of crosslinked filaments . thus the medium properties can not be set externally , but should be determined self - consistently from the properties of the test filament and its crosslinks . in particular , the studied crosslink unbinding processes reduce the connectivity , and therefore stiffness , of the medium . unbinding should , therefore , be reflected as a change in the tube potential . in the following , we will therefore assume the medium to be characterized by an energy function @xmath89 , which quantifies the energy cost of a deformation @xmath90 . the total tube potential @xmath91 then contains contributions from both the crosslinks _ and _ the effective medium . as we expect the medium to be nonlinear , @xmath92 is not necessarily a harmonic function of deformation amplitude . with a finite medium compliance , any transverse displacements , @xmath93 at a crosslink , has to be shared between the crosslink and the medium , @xmath94 . the relative stiffness of the two elements dictate the magnitude of the deformations via a force balance condition @xcite . with this the total confinement strength is no longer set by the crosslink stiffness , @xmath17 , but by an effective stiffness @xmath95 determined by the serial connection of the crosslink and the medium . unbinding events are expected to reduce the stiffness of the medium and therefore of @xmath95 . in a softer environment , however , the test filament and its crosslinks will have a _ reduced _ tendency for further crosslink unbinding . there is , therefore , a negative feedback loop between medium stiffness and crosslink unbinding . this may smooth out the sudden unbinding transition that was observed in figure [ n.energy ] . to calculate @xmath92 we need a condition of self - consistency . this is based on previous approaches @xcite . as discussed above , a macroscopic strain @xmath26 leads to a local medium deformation of @xmath96 . the associated energy cost is given by @xmath97 . at the same time the strain deforms the test filament and its surrounding tube . therefore , the energy that is needed to impose the strain @xmath26 needs to be balanced by the energy @xmath38 that is build up in the test filament . this condition can be written in the simple form @xmath98 where @xmath99 and the average is taken over crosslink occupation @xmath83 and network disorder . despite its simple form , equation ( [ eq : km.sc ] ) is rather difficult to handle , as the unknown potential @xmath92 is required for the evaluation of the statistical average . a possible solution could proceed iteratively , starting from a suitably chosen initial guess . we have not tried such a scheme . instead , and to make analytical progress , we propose a simplified ansatz for the confinement potential . as we will see , this ansatz successfully describes the intuitive result of a softening of the medium as a result of crosslink unbinding . let us assume the confinement to be harmonic in the @xmath100-degrees of freedom @xmath101 but with a confinement strength @xmath102 that depends on the deformation amplitude @xmath26 . the form of equation ( [ eq : serial_springs ] ) mimics the serial connection of crosslink and medium . the energy @xmath38 can then easily be calculated , and equation ( [ eq : km.sc ] ) takes the simple form @xmath103 which has to be solved for @xmath104 . the dependence on strain @xmath26 is implicit in the averaging procedure , as the tube potential ( [ eq : v.nonlinear.harmonic ] ) depends on @xmath26 . thus , the effective medium stiffness is obtained as a serial connection of the three mechanical elements fiber , @xmath105 , crosslinks , @xmath106 , and the medium itself , @xmath107 . figure [ graphical.sol ] presents a graphical solution of equation ( [ eq : km.sc.2 ] ) . the symbols give the effective fiber stiffness ( right - hand side of the equation , from the mc simulations ) plotted as a function of @xmath108 and taken at different deformation amplitudes @xmath26 . if we assume @xmath95 to be given and constant ( as in figure [ n.energy ] ) , a vertical line would give us the fiber stiffness as a function of amplitude @xmath26 . for @xmath95 large enough ( indicated by the vertical dashed line ) , a discontinuous transition is evident in the data . to extract the actual network modulus @xmath104 we have to find the intersection with the curve @xmath109 , equation ( [ eq : serial_springs ] ) , which is drawn as solid black line . the resulting @xmath3 is clearly decreasing with deformation amplitude , however , the discontinuous nature is not obvious anymore . a quantitative analysis is presented in figure [ km.sim.th ] . the resulting network modulus @xmath104 is plotted for different crosslink stiffness @xmath17 . for small strain the modulus @xmath3 is constant , as for a linear elastic material . this linear stiffness first increases with crosslink stiffness but then saturates , when the crosslinks become stiffer than the fiber . this behavior is in line with the serial connection between fiber , crosslink and medium as embodied in equation ( [ eq : km.sc.2 ] ) . in a serial connection it is always the softer element that governs the mechanical properties . at higher strain and for stiff crosslinks , the network modulus @xmath3 decreases . this decrease is not discontinuous as expected from figure [ n.energy ] , but rather smooth and gradual . in fact , a zero - temperature analysis along the lines of section [ sec : theor - fram ] shows that the discontinuity turns into a second - order transition , with a cusp as indicated in the right panel of figure [ km.sim.th ] . in the limit , @xmath110 , we find @xmath111 with @xmath112 and the critical amplitude @xmath113 . finally , figure [ n.energy.sc ] shows the resulting crosslink occupation @xmath83 as well as the average energy of the test filament . qualitatively , the conclusion is similar as in figure [ n.energy ] . the crosslink stiffness is identified as key factor in mediating crosslink unbinding processes . quantitatively , we see that the unbinding under load is more gradual in this second scenario , where the medium stiffness is determined self - consistently . this reflects the anticipated negative feedback of medium stiffness on crosslink unbinding . looking back at figure [ graphical.sol ] it is clear that the strength of the variation in @xmath3 or @xmath83 depends on the relative location of the two curves , @xmath109 and the effective fiber stiffness as embodied in the right - hand side of equation ( [ eq : km.sc.2 ] ) . for this geometrical factors could be important these could arise from network structural randomness , like bond angles or crosslink distances . in our calculation these geometrical factors have been disregarded , which corresponds to a regular network architecture . in this study , we discussed the interplay between filament and crosslink elasticity in semiflexible polymer networks . in particular , we were interested in the force - induced unbinding of crosslinks in response to external load . importantly , we considered the limiting case of slowly changing load ( `` quasistatic '' ) , where the system is given time enough to reach an equilibrium state at each load level . the model presented is therefore purely thermodynamic in nature , and no rate constants for the crosslink dynamics are needed . possible extensions of the present work may include the effect of a load applied at finite rates . we model the filament network as an elastic medium with modulus @xmath3 . the stiffness is calculated on a self - consistent basis from the response of a test filament that is embedded into the medium . on the microscopic level the effect of the network is to confine the filament to a tube - like region in space . we quantify the applied load in terms of a network strain @xmath26 , which is homogeneous on a macroscopic scale . on the local scale of the individual filaments , however , even a homogeneous strain leads to an inhomogeneous ( `` nonaffine '' ) deformation of the effective medium . this is a natural consequence of network heterogeneity and filament mechanical anisotropy . we modeled this inhomogeneous deformation in terms of a distortion of the center - line of the confinement tube of the filaments . the filament ( with its bending stiffness @xmath5 ) resists this tube deformation , which leads to a frustration effect between filament bending , crosslink deformation and medium deformation . this competition is formalized in equation ( [ eq : km.sc.2 ] ) which we rewrite here as @xmath114^{-1}\right\rangle_n\,.\end{aligned}\ ] ] with an effective filament bending stiffness @xmath105 , the crosslink stiffness @xmath17 and the network modulus @xmath3 . by solving this equation , the latter is thereby obtained self - consistently from a serial connection of the filament , the crosslinks and the medium itself . the most interesting feature in this equation is the dependence on the number of bound crosslinks @xmath115 . with the possibility of crosslink unbinding ( decreasing @xmath115 ) the competition between the different mechanical elements is avoided . this leads to a reduction of the medium stiffness with increasing network strain , and eventually to network failure , when all crosslinks are unbound . different scenarios can be distinguished . if the crosslinks are soft ( @xmath17 small ) , then the network modulus is dominated by the crosslinks , @xmath116 , and the filament effectively behaves as a rigid rod . the tendency for crosslink unbinding is weak as it does not lead to a significant stress relaxation . on the other hand , if crosslink and filament stiffness compete ( @xmath17 large ) , then unbinding events do help relax the imposed stress and reduce the amount of stored energy . this unbinding can be sudden and discontinuous or take the form of a second - order transition , where @xmath104 or @xmath117 display a kink at a critical load @xmath118 . experimentally , strain - induced stress relaxation is observed in living cells after transient pulses of stretch @xcite or in - vitro when the loading rates are small @xcite . for larger loading rates , a pronounced stiffening is found in the in - vitro system . this is consistent with the first - order crosslink unbinding scenario discussed here . in this picture , a free - energy barrier , and the associated time - scale , prevents crosslink unbinding when the loading rate is too large . related phenomena are important for the aging behavior of kinetically trapped actin networks @xcite , where built in stresses only relax slowly and by the action of crosslink binding events . similarly , red blood cells owe their remarkable ability to undergo reversible shape changes to a rewiring of the spectrin network @xcite . this work goes beyond previous models in considering both the filament and the crosslink stiffness as factors for the rheological properties of crosslinked filament networks . moreover , we show how this interplay affects the tendency of crosslinks to bind / unbind from the filaments during a rheological experiment . the strain field imposed by the rheometer leads to nonaffine deformations on the scale of the filaments . as compared to previous models , the unrealistic assumption of affine deformations is abandoned in favour of a model that incorporates the filament length as the fundamental non - affinity scale . in extensions of the present model one should incorporate nonlinear elastic compliances of the filaments and of the crosslinks . this is believed to be important for the nonlinear strain stiffening of f - actin networks . the complex interplay between stiffening and softening described in rheological experiments @xcite would then reflect the relative importance of filament / crosslink elasticity , which lead to stiffening , and softening as due to crosslink unbinding . interestingly , unbinding processes may under some conditions also lead to the reverse ( i.e. stiffening ) effect @xcite . this further contributes to the rich nonlinear behavior of these systems , which is far from being fully understood . support by the deutsche forschungsgemeinschaft , emmy noether program : he 6322/1 - 1 , and by the collaborative research center sfb 937 is acknowledged . 10 j. liu , g. h. koenderink , k. e. kasza , f. c. mackintosh , and d. a. weitz . visualizing the strain field in semiflexible polymer networks : strain fluctuations and nonlinear rheology of @xmath119-actin gels . , 98:198304 , may 2007 . kasza , c.p . broedersz , g.h . koenderink , y.c . lin , w. messner , e.a . millman , f. nakamura , , t.p . stossel , f.c . mackintosh , and d.a . actin filament length tunes elasticity of flexibly cross - linked actin networks . , 99:1091 , 2010 .	the mechanical properties of cells are dominated by the cytoskeleton , an interconnected network of long elastic filaments . the connections between the filaments are provided by crosslinking proteins , which constitute , next to the filaments , the second important mechanical element of the network . an important aspect of cytoskeletal assemblies is their dynamic nature , which allows remodeling in response to external cues . the reversible nature of crosslink binding is an important mechanism that underlies these dynamical processes . here , we develop a theoretical model that provides insight into how the mechanical properties of cytoskeletal networks may depend on their underlying constituting elements . we incorporate three important ingredients : nonaffine filament deformations in response to network strain ; interplay between filament and crosslink mechanical properties ; reversible crosslink ( un)binding in response to imposed stress . with this we are able to self - consistently calculate the nonlinear modulus of the network as a function of deformation amplitude and crosslink as well as filament stiffnesses . during loading crosslink unbinding processes lead to a relaxation of stress and therefore to a reduction of the network modulus and eventually to network failure , when all crosslink are unbound . this softening due to crosslink unbinding generically competes with an inherent stiffening response , which may either be due to filament or crosslink nonlinear elasticity .
the question whether interacting fermions form a landau fermi liquid state at low temperatures or rather transform into something commonly referred to as singular or non - fermi liquid is a very fundamental one . a theoretical description requires at least a qualitative understanding of the low temperature phase in order to address the relevant degrees of freedom in that phase . but for strongly interacting electronic systems this information is usually lacking , and the low - energy degrees of freedom might be very different from free electrons . however , the fermi liquid picture is often appropriate at low temperatures and energies . reasons for this can be found in the stability of the fermi liquid fixed point of the renormalization group analysis@xcite or phase space arguments.@xcite in both approaches , the analyticity of the scattering vertices is crucial for the arguments . in situations , where the fermi liquid description fails , the origin for this is linked to low - dimensionality of the system , long - range interactions , nested fermi surfaces or the proximity to a quantum critical point , just to name a few ( see , e.g. refs . ) . the quasiparticle concept might still be valid , but the scattering rate is strongly enhanced and shows an unusual , i.e. non - quadratic , temperature and frequency dependence . alternatively the quasiparticle concept may also break down completely . in this work we will focus on the single - band hubbard model which is a minimal model for strongly correlated electron systems , and obtain a non - perturbative approximative solution by means of the dynamical mean - field theory ( dmft).@xcite within dmft , where only local correlations are included , the generic low temperature state of the metallic hubbard model is a fermi liquid.@xcite this is expected , since the low temperature phase of the spin-@xmath1 single impurity anderson model the model onto which the hubbard model is mapped within dmft is a local fermi liquid under the assumption of a `` well behaved '' medium.@xcite the fermi liquid reveals itself via a low - energy many - body bandstructure forming around the fermi level at low temperatures.@xcite due to the neglect of momentum dependent correlations in dmft the heavy quasiparticles possess the non - interacting fermi surface.@xcite a characteristic low - energy scale @xmath0 is associated with this lattice version of the kondo effect and marks the temperature scale for local - moment screening and the emergence of coherent quasiparticles . on the other hand it is known from perturbation theory and renormalization group treatments , that a van hove singularity of the non - interacting density of states ( dos ) in the vicinity of the fermi level lead to a marginal@xcite or non - fermi liquid form of the self - energy.@xcite this should be contrasted to nested fermi liquids,@xcite where due to the nesting property of the fermi surface the phase space volume for low - energy scattering is also strongly enhanced and unusual low - energy properties emerge as well.@xcite van hove singularities are a consequence of maxima or saddle points in the dispersion relation @xmath2 . for periodic energy bands in a crystal a certain number of these van hove singularities must occur on topological grounds ( see , e.g. ref . ) . for example , all of the cubic lattices have noticeable van hove singularities in space dimensions @xmath3 , typically logarithmic divergences or square - root cusps . anomalous low - energy behavior can already be anticipated by recognizing , that the usual arguments of microscopic fermi liquid theory break down if a maximum or saddle point of the dispersion relation @xmath2 is found at the fermi level . then , @xmath2 _ can not _ be approximated by a linear dispersion , i.e.@xmath4 , with @xmath5 and @xmath6 the fermi wave vector . around these points a quadratic ( maximum ) or hyperbolic ( saddle - point ) functional dependence @xmath7 results , where @xmath8 and @xmath9 are two linear independent directions in momentum space . the arguments leading to a quadratic fermi liquid energy and temperature dependence of the scattering rate , i.e. im@xmath10 , are thus not applicable . instead , the phase space volume for scattering is strongly enhanced and anomalous @xmath11- and @xmath12-dependencies may result . if the singularity occurs in the vicinity of the fermi level , strong modifications of the low - energy and temperature properties are still expected due to a pinning effect.@xcite in the present work we show that the non - fermi liquid signatures which were encountered in previous studies@xcite of the self - energy and transport properties can also occur in the paramagnetic phase of a strongly correlated metal within dmft . we use the two - dimensional square lattice with nearest and next - nearest neighbor hopping , @xmath13 and @xmath14 respectively . for @xmath15 the dispersion relation @xmath2 has saddles at the @xmath16 points in the brillouin zone ( @xmath17 ) . this leads to a logarithmic divergence in the non - interacting dos , which is located at the fermi level @xmath18 for @xmath19 and moved to finite energies for @xmath20 . the resulting anomalies consist of an unusual maximum in the imaginary part of the self - energy im@xmath21 in the vicinity of the fermi level at low but finite temperatures . the concept of quasiparticles remains valid but they are strongly scattered at the flat parts of the dispersion relation and the fermi liquid formation is disturbed . the anomalous double well structure persists over a range of doping and next - nearest neighbor hopping @xmath14 . at zero temperature and for half - filled square lattice the double - well structure is removed and the self - energy vanishes at the fermi level . even though the self - energy exhibits a non - fermi liquid energy dependence , the quasiparticle weight is still finite . the pinning of the van hove singularity to the fermi level , which has previously been found,@xcite is explicitly observed in the correlated metal . qualitatively similar non - fermi liquid signatures are also observed in more advanced theories which include nonlocal correlations.@xcite these features are usually attributed to the presence of strong nonlocal antiferromagnetic correlations . however , in the present approach nonlocal correlations are not included and we argue that an interpretation _ exclusively _ in terms of nonlocal correlations is too simplified . the paper is organized as follows . in section [ sec : theo ] we shortly present the hubbard hamiltonian and the basic equations of dmft . we elucidate the role of the non - analyticities in the non - interacting dos by means of a simplified model dos for which analytical insights are obtained in sect . [ sec : hubnflcusp ] . self - consistent dmft calculations for singular and non - singular model dos reveal the non - analyticity to be the cause for the observed non - fermi liquid signatures . in sect . [ sec : square ] we present finite and zero temperature calculations for the spectral function and self - energy for a square lattice without next - nearest neighbor hopping @xmath19 . the anomalies are shown to occur at temperatures on the order of the low energy scale @xmath0 and decrease with temperature . at @xmath22 a non - quadratic energy dependence in the imaginary part of the self - energy remains . the case of the square lattice with next - nearest neighbor hopping @xmath23 is discussed in sect . [ sec : ttp ] . in addition to the spectral function and the self - energy , we also study the temperature dependence of the zero frequency quasiparticle scattering rate and the resistivity . we discuss in sect . [ sec : susnflrelpg ] the relevance of the present findings to the cuprate superconductors before we conclude the paper . we study the single band hubbard model @xmath24 where the operator @xmath25 ( @xmath26 ) annihilates ( creates ) an electron in a localized wannier orbital at lattice site @xmath27 with spin @xmath28 , @xmath29 represents the number operator for electrons , @xmath30 is the local ionic level position at each site , and @xmath31 is the local matrix element of the coulomb repulsion . in the following an external magnetic field is not included and all spin up and down quantities are equal . the spin label of green s functions and self - energies will be suppressed . we focus on the paramagnetic phase of the model and exclude possible phase transitions to ordered phases such as superconductivity or magnetism . all information on the lattice is encoded in the one - particle hopping amplitude @xmath32 . its fourier - transform gives the dispersion relation @xmath33 where we set the lattice constant @xmath34 and already specified to the two - dimensional ( @xmath35 ) square lattice with nearest and next - nearest neighbor hopping , @xmath13 and @xmath14 , respectively . the central quantity for the discussion of fermi liquid properties is the correlation self - energy @xmath36 . the assumption that it does not depend on momentum @xmath37 constitutes the major approximation of dmft . the lattice green s function of the hubbard model is then given by @xmath38 which only depends on @xmath37 via the dispersion relation @xmath2 . the latter also determines the non - interacting dos @xmath39 . within dmft the lattice model is mapped onto an effective impurity model , where the medium @xmath40 of the impurity model can not be chosen arbitrarily , but has to be determined from the self - consistency conditions @xmath41 the functional form of the local scattering matrix @xmath42 is completely determined by the non - interacting dos . the correlation self - energy enters only as an unspecified parameter [ via @xmath43 and is obtained from the solution of the effective impurity model for a given medium @xmath40 . the self - consistency cycle is closed by identifying the local green s function of the lattice with the one of the impurity model , @xmath44 for the finite temperature calculations we use the enhanced non - crossing approximation@xcite ( enca ) as the impurity solver . there are no adjustable parameters and the method works directly on the real frequency axis . this is crucial for the observation of the small low - energy anomalies in the self - energy . however , it can not be employed at too low temperatures @xmath45 due to the nca pathology.@xcite within the enca the pathologic behavior is considerably improved compared to nca.@xcite it can be reliably used down to temperatures @xmath46 for moderate to large values of @xmath31 and not too far away from half - filling . at zero temperature we use the numerical renormalization group@xcite ( nrg ) instead . for the nrg spectral functions we use the complete fock - space@xcite algorithm of ref . with a discretization parameter @xmath47 and keep approximately @xmath48 states in each nrg - iteration step . we average over eight different discretizations@xcite of the conduction band to minimize the errors . we also performed finite temperature dmft(nrg ) calculations which were in accord with the dmft(enca ) results and confirmed the existence of the anomalous structures in the self - energy . however , the anomalous features occur at energies and temperatures where the results depend on the actual values of nrg broadening parameters.@xcite therefore , we prefer the enca for the finite temperature calculations . before presenting dmft results for the hubbard model on cubic lattices we elucidate the basic mechanism by means of a simplified model dos of the form @xmath49 , \quad \alpha\geq 0\end{aligned}\ ] ] with @xmath50 . for @xmath51 this dos has a non - analytic cusp at @xmath18 which turns into a logarithmic divergence for @xmath52 , @xmath53 , while it is smooth for @xmath54 ( see inset of fig . [ fig : hubtoydosttilde ] ) . the case @xmath55 mimics the logarithmic divergence of the @xmath35 simple cubic lattice . the explicit form ( [ eq : hubtoydoscusp ] ) allows for the analytic calculation of the effective medium @xmath40 as shown in the appendix . we state the result for three values of @xmath56 @xmath57 - \text{li}_2[\tilde{g}(z)w ] } \ ] ] @xmath58 + 2\mathrm{atanh}[\tilde{g}(z ) w ] } \ ] ] @xmath59 } \quad,\end{aligned}\ ] ] where li@xmath60 is the dilogarithm . the crucial difference in the analytic structure of the three results is the appearance of non - analytic logarithms for @xmath55 and @xmath61 which are absent for @xmath62 . suppose the system is about to form a fermi liquid state as it is expected within dmft for the hubbard model at low @xmath11 . then , the correlation self - energy has the form , @xmath63\right ) , \end{aligned}\ ] ] where @xmath11 is the temperature , @xmath64 a measure for the residual quasiparticle scattering , and @xmath65 the quasiparticle weight . the renormalized single - particle energy @xmath66 determines the fermi wave vector via @xmath67 and vanishes in symmetric situations . using the form ( [ eq : flsig ] ) in the expressions eqs . ( [ eq : hubtoylindos0 ] ) to ( [ eq : hubtoylindos2 ] ) , the imaginary part of @xmath68 develops a cusp - like minimum at the fermi level for small @xmath64 and @xmath69.@xcite figure [ fig : hubtoydosttilde ] illustrates this for the three different values of the exponent @xmath70 , and @xmath71 . [ see eqs . , and ] as functions of energy @xmath72 for the cases @xmath73 and two values of the residual interaction @xmath64 . the inset shows the model dos for the same values of @xmath56 . @xmath74 and @xmath75 was used . , width=302 ] in the limit @xmath76 the effective media for @xmath61 and @xmath62 approach the values im@xmath77 and im@xmath78 , respectively , while for @xmath55 the effective medium eventually reaches zero ( see eqs . ( [ eq : tta0w0 ] ) to ( [ eq : tta2w0 ] ) ) . thus , for @xmath55 , we end up with a soft - gap effective medium for which kondo screening and the local fermi liquid behavior can be completely destroyed under certain conditions.@xcite however , in the present case the effective medium vanishes only logarithmically at the fermi level , i.e.im@xmath79 [ see eq . ( [ eq : tta0w0 ] ) ] , and therefore the low temperature properties are still expected to be characterized by the strong coupling fixed point as will be shown in sect . [ sec : nrg ] . the topology of the fermi surface does not enter the present arguments at any point , only the non - analyticity in the non - interacting dos does . the latter derives from flat parts in the dispersion relation , which in the case of the @xmath35 square lattice are saddle points at the @xmath16 points . it is therefore clear , that no nesting property of the fermi surface , which is present in the square lattice at half - filling , can be responsible for these effects . in all calculations of this work the hopping matrix element @xmath13 is used as the energy scale and energies are measured in units @xmath80 , setting the effective half bandwidth @xmath81 . in this section the next - nearest neighbor hopping is set to zero , @xmath19 . the non - interacting dos of this @xmath35 square lattice has the same logarithmic singularity at @xmath18 as the model dos of eq . ( [ eq : hubtoydoscusp ] ) for @xmath55 . results of a dmft(enca ) calculation for the half - filled @xmath82 hubbard model on a square lattice are shown in fig . [ fig : hubb_2dsc_sym ] for various temperatures . for the symmetric hubbard model on a @xmath35 square lattice for coulomb repulsion @xmath83 ( @xmath84 ) , half bandwidth @xmath85 ( @xmath86 ) and various temperatures as functions of frequency within dmft(enca ) . the thin dashed curve represents the non - interacting dos @xmath87 , and the inset shows the imaginary part of the effective medium im@xmath88 . ( b ) imaginary ( left ) and real ( right ) part of the self - energy @xmath21 for the same parameters as in ( a ) and @xmath11 as indicated . the inset of the left panel shows the imaginary part on a wider energy interval . the short arrows in the right graphs point at the approximate positions of the kinks . , title="fig:",width=302 ] for the symmetric hubbard model on a @xmath35 square lattice for coulomb repulsion @xmath83 ( @xmath84 ) , half bandwidth @xmath85 ( @xmath86 ) and various temperatures as functions of frequency within dmft(enca ) . the thin dashed curve represents the non - interacting dos @xmath87 , and the inset shows the imaginary part of the effective medium im@xmath88 . ( b ) imaginary ( left ) and real ( right ) part of the self - energy @xmath21 for the same parameters as in ( a ) and @xmath11 as indicated . the inset of the left panel shows the imaginary part on a wider energy interval . the short arrows in the right graphs point at the approximate positions of the kinks . , title="fig:",width=302 ] in part ( a ) the non - interacting dos as well as the fully interacting local spectral function @xmath89 are displayed for @xmath90 . the well separated hubbard bands around the ionic level positions @xmath91 are clearly visible . upon lowering the temperature the lattice version of the kondo - effect leads to the build - up of a pronounced many - body resonance at the fermi level . the van hove singularity in @xmath92 leads to the unusual minimum in the imaginary part of the effective medium im@xmath88 [ see inset of fig [ fig : hubb_2dsc_sym](a ) ] as it was already found in the previous section . in contrast to fig . [ fig : hubtoydosttilde ] the minimum in im@xmath88 is not cusp - like at @xmath18 but smeared out due to the imaginary part of the self - energy which is shown in part ( b ) of the figure . however , the non - analytic structure in the lattice scattering matrix , eqs . ( [ eq : scatter ] ) and ( [ eq : hubtoylindos0 ] ) , is sufficient to produce this minimum at low temperatures . the self - consistency of dmft leads to an accompanying maximum in im@xmath21 at @xmath18 , which is in striking contrast to fermi - liquid behavior im@xmath10 for small @xmath72 and @xmath11 . the position of the two minima emerging at finite frequencies next to the central maximum is moved toward the fermi level with decreasing temperature so that the anomalous region shrinks . as a side remark we add here , that we observe additional kinks in the real part of the self - energy as shown in fig . [ fig : hubb_2dsc_sym](b ) , which are associated with this non - fermi liquid behavior . at energies @xmath93 the known kinks arising from the coupling of the quasiparticles to local spin fluctuations are observable.@xcite the additional kinks are clearly visible at lower energies . in contrast to the kinks at higher energy , their position is temperature dependent and moves toward the fermi level with decreasing @xmath11 . anomalous maxima in im@xmath94 are also encountered within the functional renormalization group approach to the hubbard model at finite temperatures.@xcite crucial for these anomalies to occur is the renormalization of the two - particle interaction vertex . within dmft all _ local _ @xmath95-particle vertices and their renormalizations are included . in im@xmath94 in units of the low - energy scale @xmath0 as a function of the rescaled temperature @xmath96 for three different coulomb repulsions . the inset shows the screened local magnetic moment normalized to its high temperature value @xmath97 as a function of temperature for the same values of the coulomb repulsions . the horizontal line indicates the value where @xmath98 reaches @xmath99 of @xmath97 and the vertical dashed lines indicate the low - energy scales : @xmath100 , @xmath101 , and @xmath102 . , width=302 ] we extract the low - energy scale @xmath0 as the temperature where the effective screened local moment as calculated from the local magnetic susceptibility via@xcite @xmath103 is reduced to @xmath99 of its high temperature value @xmath97 ( see inset of fig . [ fig : hub_2dsc_scale ] ) . this yields the equivalent result to the width of the many - body resonance at the fermi level in the momentum resolved spectral function im@xmath104 ( not shown ) . the temperature evolution of the anomalous double - well structure is governed by the low - energy scale as revealed by fig . [ fig : hub_2dsc_scale ] . there , the position of the anomalous minimum @xmath105 in im@xmath94 as a function of temperature for three values of the coulomb repulsion is shown ( the self - energies as well as the minima at @xmath106 are symmetric around @xmath18 for half - filling ) . for all @xmath31 , @xmath107 exhibits a dome - shaped curve with both , maximum position and height determined by @xmath0 . the physical origin can therefore be directly linked to the emergence of the low - energy quasiparticles . at high temperatures the system is dominated by the incoherent charge and spin excitations and the van hove singularity is concealed . at temperatures of the order of @xmath0 the lattice version of the kondo effect leads to the screening of local magnetic moments . the break - up of screened magnetic moments gives rise to long - lived low - energy quasiparticle excitations which manifest themselves in the many - body resonance at the fermi level as observed in the spectral function of e.g. fig . [ fig : hubb_2dsc_sym](a ) . but their evolution into the coherent fermi liquid quasiparticle is disturbed by the enhanced scattering due to the van hove singularity . this leads to the maximum in im@xmath94 at the fermi level . further lowering the temperature the phase space volume for quasiparticle scattering shrinks and so does the extension of the maximum . the van hove singularity is shifted away from the fermi energy @xmath18 when doping the system , @xmath108 . the corresponding anomalous minimum in the effective medium im@xmath68 and the maximum in im@xmath94 are too moved to finite energies . this can be seen in fig . [ fig : hubb_gf_2dsc_asym ] , where the spectral function , the self - energy , and the effective medium are shown for the hubbard model on a @xmath35 square lattice ( @xmath19 ) with @xmath109 and @xmath110 ( filling @xmath111 ) . square lattice with @xmath109 and filling @xmath111 for various temperatures obtained with the dmft(enca ) . the inset shows the low - energy region . ( b ) imaginary part of the self - energy around the fermi level for the same parameters as in ( a ) and temperatures as indicated . the inset shows the imaginary part of the effective medium in the low - energy region . , title="fig:",width=302 ] square lattice with @xmath109 and filling @xmath111 for various temperatures obtained with the dmft(enca ) . the inset shows the low - energy region . ( b ) imaginary part of the self - energy around the fermi level for the same parameters as in ( a ) and temperatures as indicated . the inset shows the imaginary part of the effective medium in the low - energy region . , title="fig:",width=302 ] the quasiparticle peak around @xmath18 in the spectral function , as displayed in the inset of fig . [ fig : hubb_gf_2dsc_asym](a ) , acquires a pronounced asymmetry due to the double - well structure in im@xmath94 . the shoulder in the curve for the lowest temperature might even be interpreted as a precursor of a pseudogap . the flat parts in the dispersion relation are located energies away from the fermi level and the dispersion relation can again be approximated linearly in a vary narrow region around the fermi surface . in this region the usual arguments of microscopic fermi liquid theory hold and the system forms a regular fermi liquid at very low temperatures . this is observed in the self - energy shown in fig . [ fig : hubb_gf_2dsc_asym](b ) where a quadratic minimum forms at @xmath18 at @xmath112 . however , the fermi liquid parameters will be strongly renormalized as compared to a situation without the van hove singularity . + + no true divergence in the non - interacting dos is necessary for the anomalies described above to occur . a non - analytic cusp as in the model dos of eq . ( [ eq : hubtoydoscusp ] ) for @xmath113 is sufficient . this is illustrated in fig . [ fig : hubrealdosrunsttsg ] where fully self - consistent dmft calculations for several model dos are considered . the first two plots ( a ) and ( b ) are for @xmath114 with @xmath61 at and away from half - filling , respectively . figures ( c ) and ( d ) are for asymmetric dos which are constant below the fermi level and have linear ( c ) and a square - root ( d ) cusp at @xmath18 . the imaginary parts of the effective media and self - energies show the characteristic minima and maxima induced by the non - analyticities in @xmath92 . for the half - filled cases ( a ) , ( c ) and ( d ) the anomalies are situated at the fermi level , while for finite hole doping as in ( b ) these are moved to positive energies . figure [ fig : hub - nrg_2d_sc_sym](a ) displays the spectral function for the hubbard model on a square lattice with @xmath83 at zero temperature and for different doping @xmath115 . these spectra were obtained with the nrg@xcite as impurity solver . for the hubbard model on a @xmath35 square lattice for @xmath83 and varying hole doping . the inset shows the position of the maximum as a function of doping for @xmath83 and @xmath116 . the dashed line is a fit @xmath117 . ( b ) imaginary part of the self - energy at half - filling @xmath118 for three values of the coulomb repulsion @xmath31 . the curves for @xmath119 and @xmath120 were rescaled in order to lie on top of the @xmath83 curves at low @xmath72 . the dashed line is a fit with a function @xmath121 . the upper inset shows the low energy behavior in a double - logarithmic plot . the lower inset shows the quasiparticle weight for @xmath83 as a function of doping . , title="fig:",width=302 ] for the hubbard model on a @xmath35 square lattice for @xmath83 and varying hole doping . the inset shows the position of the maximum as a function of doping for @xmath83 and @xmath116 . the dashed line is a fit @xmath117 . ( b ) imaginary part of the self - energy at half - filling @xmath118 for three values of the coulomb repulsion @xmath31 . the curves for @xmath119 and @xmath120 were rescaled in order to lie on top of the @xmath83 curves at low @xmath72 . the dashed line is a fit with a function @xmath121 . the upper inset shows the low energy behavior in a double - logarithmic plot . the lower inset shows the quasiparticle weight for @xmath83 as a function of doping . , title="fig:",width=302 ] as the hole - doping is increased , the lower hubbard band moves toward the fermi level and eventually merges with the quasiparticle peak . for very large doping the lower hubbard band is moved above the fermi level and @xmath122 resembles the non - interacting dos as the system becomes effectively non - interacting . the position of the quasiparticle peak is attracted to the fermi level up to considerable values of doping ( roughly @xmath123 ) . this pinning becomes more visible in the inset , where the position @xmath124 of the maximum in @xmath125 is plotted as a function of doping @xmath115 for @xmath83 and @xmath116 . in the interacting case @xmath124 depends quadratically on the doping , @xmath126 , and is considerably reduced compared to @xmath116 . the many - body correlations renormalize the van hove singularity toward the fermi level especially at small doping @xmath123 . the imaginary part of the self - energy shown in fig . [ fig : hub - nrg_2d_sc_sym](b ) does not display any anomalous double - well structure . the logarithmic van hove singularity in @xmath92 is a rather weak divergence and at zero temperature the usual arguments leading to a vanishing self - energy at the fermi level apply.@xcite however , im@xmath94 shows an anomalous frequency dependence at low energies , as it grows like im@xmath127 which was also previously found in perturbation theory.@xcite is found in perturbation theory for quasiparticles at the fermi level , but not at the van hove singularity , which scatter from quasiparticles at the van hove singularities . for quasiparticles at the van hove singularity im@xmath128 and scattering processes not involving the saddle points produce the usual @xmath35 fermi liquid result im@xmath129 . dmft does not resolve the different momentum dependencies in the interaction channels and averages over the whole brillouin zone . due to its large phase space , the contribution @xmath130 dominates this average and is therefore observed in the dmft . this is confirmed by fitting the imaginary part with a function @xmath131 which yields a very small linear contribution @xmath132 without changing the range of validity of the fit . the usual @xmath35 fermi contribution @xmath133 vanishes too fast for @xmath134 to make a significant contribution at low energies . ] the upper inset reveals this low energy behavior more clearly in a double logarithmic plot . the quasiparticle peak at the fermi level for half - filling @xmath118 shows a logarithmic divergence as was already observed recently.@xcite but in contrast to what was stated in reference , the quasiparticle weight @xmath135 does not vanish at half - filling in our calculation , as can be seen in the inset of fig . [ fig : hub - nrg_2d_sc_sym](b ) . ( the quasiparticle weight in fig . 3 of ref . also seems to extrapolate to a finite value at half - filling but a contradicting statement is made in the text of that reference . ) the slope of the real part of the self - energy does consequently not diverge at the fermi level . this is in accord with the low frequency dependency of the imaginary part @xmath136 from which the real part is obtained via kramers - kronig relation , re@xmath137 $ ] . the system is therefore well characterized by a generalized fermi liquid@xcite where im@xmath94 shows some anomalous frequency dependence , but still vanishes at the fermi level and the quasiparticle weight remains finite . ( left panels ) and imaginary parts of the self - energy @xmath21 ( right panels ) for different filling @xmath95 for the hubbard model on the square lattice with additional next - nearest neighbor hopping @xmath138 and @xmath139 . the insets in the left panels show the imaginary part of the effective medium while the insets on the right show a close - up of the low - energy region of im@xmath94 . the non - interacting dos is shown in the inset for the half - filled case ( b ) . , title="fig:",width=170 ] ( left panels ) and imaginary parts of the self - energy @xmath21 ( right panels ) for different filling @xmath95 for the hubbard model on the square lattice with additional next - nearest neighbor hopping @xmath138 and @xmath139 . the insets in the left panels show the imaginary part of the effective medium while the insets on the right show a close - up of the low - energy region of im@xmath94 . the non - interacting dos is shown in the inset for the half - filled case ( b ) . , title="fig:",width=170 ] ( left panels ) and imaginary parts of the self - energy @xmath21 ( right panels ) for different filling @xmath95 for the hubbard model on the square lattice with additional next - nearest neighbor hopping @xmath138 and @xmath139 . the insets in the left panels show the imaginary part of the effective medium while the insets on the right show a close - up of the low - energy region of im@xmath94 . the non - interacting dos is shown in the inset for the half - filled case ( b ) . , title="fig:",width=170 ] ( left panels ) and imaginary parts of the self - energy @xmath21 ( right panels ) for different filling @xmath95 for the hubbard model on the square lattice with additional next - nearest neighbor hopping @xmath138 and @xmath139 . the insets in the left panels show the imaginary part of the effective medium while the insets on the right show a close - up of the low - energy region of im@xmath94 . the non - interacting dos is shown in the inset for the half - filled case ( b ) . , title="fig:",width=170 ] ( left panels ) and imaginary parts of the self - energy @xmath21 ( right panels ) for different filling @xmath95 for the hubbard model on the square lattice with additional next - nearest neighbor hopping @xmath138 and @xmath139 . the insets in the left panels show the imaginary part of the effective medium while the insets on the right show a close - up of the low - energy region of im@xmath94 . the non - interacting dos is shown in the inset for the half - filled case ( b ) . , title="fig:",width=170 ] ( left panels ) and imaginary parts of the self - energy @xmath21 ( right panels ) for different filling @xmath95 for the hubbard model on the square lattice with additional next - nearest neighbor hopping @xmath138 and @xmath139 . the insets in the left panels show the imaginary part of the effective medium while the insets on the right show a close - up of the low - energy region of im@xmath94 . the non - interacting dos is shown in the inset for the half - filled case ( b ) . , title="fig:",width=170 ] ( left panels ) and imaginary parts of the self - energy @xmath21 ( right panels ) for different filling @xmath95 for the hubbard model on the square lattice with additional next - nearest neighbor hopping @xmath138 and @xmath139 . the insets in the left panels show the imaginary part of the effective medium while the insets on the right show a close - up of the low - energy region of im@xmath94 . the non - interacting dos is shown in the inset for the half - filled case ( b ) . , title="fig:",width=170 ] ( left panels ) and imaginary parts of the self - energy @xmath21 ( right panels ) for different filling @xmath95 for the hubbard model on the square lattice with additional next - nearest neighbor hopping @xmath138 and @xmath139 . the insets in the left panels show the imaginary part of the effective medium while the insets on the right show a close - up of the low - energy region of im@xmath94 . the non - interacting dos is shown in the inset for the half - filled case ( b ) . , title="fig:",width=170 ] in this section we extend the study to an additional next - nearest neighbor hopping @xmath23 . the local spectral functions and self - energies displayed in fig . [ fig : hubb_2dscnnn_asym ] for @xmath139 , various fillings and temperatures show pronounced particle - hole asymmetries . at low temperatures the anomalous minima in im@xmath68 and corresponding maxima in the self - energy im@xmath94 can be observed slightly below the fermi level for filling @xmath140 , @xmath82 , and @xmath111 . this correlates with the logarithmic van hove singularity in @xmath92 shifted below the fermi level , as displayed in the right inset of fig . [ fig : hubb_2dscnnn_asym](b ) . for decreasing filling ( increasing hole doping ) the van hove singularity is moved toward the fermi level and the anomalous structures in im@xmath94 and im@xmath68 narrow . for the largest hole doping @xmath141 the imaginary part of the self - energy exhibits only one narrow minimum at very low temperature . within dmft the fermi surface does not change upon increasing the interaction @xmath31,@xcite due to the momentum independence of the self - energy . the filling at which the van hove singularity moves across the fermi level can be calculated with the non - interacting dos to be @xmath142 , which is much less than @xmath141 therefore a double - well structure would be expected but is not observed . the reason is found in the lower hubbard band reaching the fermi level . a growing number of empty lattice sites are created and disturb the quasiparticle formation . additional quasiparticle scattering results and conceals the effect of the van hove singularity . due to the pinning of the singularity to the neighborhood of the fermi level its influence is strong for all values of filling . the anomalies correlate with the low - energy scale and occur at temperatures following the trend given by @xmath0 as function of filling @xmath95 . @xmath0 reaches a minimum for @xmath143 to @xmath144 where correlations are strongest and the system is closest to the mott insulator . the anomalous behavior found in the self - energy affects the temperature dependence of transport properties . the static quasiparticle scattering rate at the fermi level im@xmath145 is shown in fig . [ fig : hubb_2dscnnn_asymtrans](a ) as a function of temperature for different filling . for the hubbard with @xmath139 on a square lattice with next - nearest neighbor hopping @xmath23 as function of temperature in log - log plot . the curves for different filling are multiplied by the factors indicated on the right . ( b ) resistivity @xmath146 ( normalized to @xmath147 ) for the same parameters as in ( a ) . the inset shows @xmath146 without the scaling factors and in a non - logarithmic plot . the dashed lines represent fits with the indicated functions in both plots . , title="fig:",width=302 ] for the hubbard with @xmath139 on a square lattice with next - nearest neighbor hopping @xmath23 as function of temperature in log - log plot . the curves for different filling are multiplied by the factors indicated on the right . ( b ) resistivity @xmath146 ( normalized to @xmath147 ) for the same parameters as in ( a ) . the inset shows @xmath146 without the scaling factors and in a non - logarithmic plot . the dashed lines represent fits with the indicated functions in both plots . , title="fig:",width=302 ] for a @xmath35 fermi liquid the expected temperature dependence of the scattering rate is given by im@xmath148 . this behavior is found for low temperatures at half - filling [ lowest curve in fig . [ fig : hubb_2dscnnn_asymtrans](a ) ] and above , i.e. @xmath149 ( not shown ) . hole - doping induces qualitative changes in the low temperature behavior since the van hove singularity is moved toward the fermi level . for filling @xmath111 and @xmath150 power - law behavior with an exponent close to one is observed , im@xmath151 . for large hole doping , @xmath141 , the scattering rate is again found to be in accord with usual fermi liquid variation , @xmath152 . this is expected to hold for even larger hole - doping @xmath153 , but could not be checked within the present approach due to restriction of the enca to relatively high temperatures at large doping . we want to emphasis that the van hove singularity induces qualitative changes in the asymptotic low temperature behavior of the scattering rate when moved closer to the fermi surface by doping . im@xmath154 decreases much slower with temperature for fillings around @xmath155 when the double - well structure is found in the vicinity of the fermi level . with the self - energy at hand one can also calculate transport coefficients . no vertex corrections occur in dmft due to the momentum independence of the self - energy and two - particle vertex.@xcite the current - current correlation function is completely determined by the particle - hole bubble . transport quantities can be expressed in terms of a generalized transport lifetime ( see , e.g. , ref . or ) @xmath156^\alpha\:{\omega}^{\beta-1}\end{aligned}\ ] ] where @xmath157 is the fermi function . the generalized transport lifetime is given by @xmath158 ^ 2 \\ = & \frac{1}{\pi^2}\int\!d\epsilon\ : \tilde\rho_0(\epsilon)\left[\text{im}g(\epsilon,{\omega}-i0^+)\right]^2 \quad.\end{aligned}\ ] ] the function @xmath159 can be calculated for the cubic lattices and turns out to be a smooth function where the singularities of @xmath92 are removed due to the derivative . only in the fermi liquid regime is @xmath160 and the linearized boltzmann theory is recovered.@xcite the resistivity @xmath161 with @xmath162 ( @xmath163 is the lattice spacing , @xmath164 the electronic charge ) is shown in fig . [ fig : hubb_2dscnnn_asymtrans](b ) for different filling . typical characteristics of strongly correlated systems can be observed for not too large doping . the resistivity @xmath146 changes from insulating ( or semi - conducting ) behavior at high temperatures to metallic behavior at low temperatures due to the emergence of ( coherent ) low - energy quasiparticles ( see inset).@xcite the high - temperature resistivity is much larger than the value expected from the ioffe - regel condition valid for usual metals , where @xmath146 is bound by a minimal scattering length of the order of the lattice spacing.@xcite at intermediate temperatures @xmath165 linear regimes with increasing slope for decreasing doping can be recognized , in accord with earlier studies.@xcite the log - log plot ( main panel ) reveals that the low temperature fermi liquid behavior @xmath166 is found for @xmath82 and above ( not shown ) . and not the two - dimensional form @xmath167 is not completely clear to us at present . it is probably related to the neglect of vertex correction within dmft . ignoring the momentum dependence in the two - particle vertex leads to a violation of momentum conservation at internal vertices.@xcite the resultant mean - field decoupling in the bethe - salpeter equations@xcite for the two - particle current - current correlation function introduces additional averages over the brillouin zone and mimics an effectively higher dimensionality for transport quantities . ] a @xmath168-dependency is also observed for @xmath141 . but for the intermediate values @xmath111 and @xmath150 a qualitative different behavior is found . @xmath169 more closely resembles power - laws with exponents @xmath144 and @xmath170 , respectively . the van hove singularity is close to the fermi level and the enhanced scattering increases the resistivity . in summary , the interplay between the reduction of phase space volume for scattering and the qualitative changes in the frequency dependence of the self - energy within the temperature window causes the qualitative changes in the transport properties . the @xmath13-@xmath14-hubbard model might be suitable to describe the low - energy physics of cuprate superconductors . even though dmft is non - perturbative and includes nontrivial _ local _ many - body correlations , it represents a rather poor approximation for the two - dimensional cuprate layers due to the neglect of _ nonlocal _ correlations . therefore , we want to point out that we neither try to model the cuprate superconductors nor can our results be directly transferred to those systems . however , even in more appropriate theories like cluster - dmft@xcite the van hove singularities associated with the lattice structure are retained and therefore the findings of this work have some bearings on the pseudogap and strange metal phase of cuprate superconductors . in the following we will elaborate on some of these aspects . the van hove singularity provides an additional mechanism for kinks@xcite in the real part of the self - energy [ see fig . [ fig : hubb_2dsc_sym](b ) ] . the location of the kinks is temperature dependent which should make it possible to discriminate these type of kinks in experiment or calculations . and @xmath139 as function of temperature for three different filling obtained from dmft(enca ) calculations and eq . ( [ eq : qpztf2 ] ) . , width=226 ] the non - fermi liquid self - energy induces an unusual temperature dependence of the generalized quasiparticle weight@xcite @xmath171 where @xmath172 is the smallest fermionic matsubara frequency ( the form ( [ eq : qpztf1 ] ) is useful when the self - energy is obtained only at the imaginary matsubara frequencies like in quantum monte - carlo calculations ) . the generalized quasiparticle weight as calculated with eq . ( [ eq : qpztf2 ] ) is shown in fig . [ fig : qpz ] for the hubbard model on the square lattice with next - nearest neighbor hopping as function of temperature and three different filling . all curves show a change in curvature at the temperature where the anomalous double - well structure emerges . this behavior closely resembles that found in ref . for a crossover from a marginal fermi liquid at high temperatures to a fermi liquid at low @xmath11 . since the self - energies do not show any signs of marginal fermi liquid behavior ( see the previous sections ) this reveals a difficulty when interpreting imaginary matsubara data at finite temperatures . due to the integral in eq . ( [ eq : qpztf2 ] ) the temperature dependence of @xmath173 is not determined by the asymptotic low - frequency form of im@xmath21 but by its average spectral weight in the temperature window @xmath174 . therefore , the discrimination between different characteristic low - energy forms of @xmath21 given the self - energy at imaginary matsubara frequencies seems rather delicate at finite temperature . the partial destruction of the fermi surface ( fermi arcs ) and the emergence of a pseudogap@xcite is of current interest . these phenomena are usually explained to arises from the coupling to strong nonlocal antiferromagnetic correlations.@xcite in the present study no nonlocal fluctuations are included but instead the van hove singularity produces similar features such as a non - fermi liquid maximum in im@xmath175 and the concomitant reduction of spectral weight [ see fig . [ fig : hubb_gf_2dsc_asym](a ) ] , however , a true pseudogap is not observed in our approach indicating the importance of non - local fluctuations . the existence of fermi arcs and the asymmetry for momentum vectors along the nodal @xmath176 ( @xmath177 ) and anti - nodal @xmath178 direction is natural in the present scenario . the flat parts in the dispersion relation occur at the @xmath16 point in the brillouin zone where the low - temperature quasiparticles are most strongly affected and consequently the fermi liquid description breaks down there first.@xcite also the stability of the induced non - fermi liquid behavior up to a doping of the order of @xmath179 which was encountered in other studies@xcite can be explained by the pinning of the van hove singularity to the fermi level . unarguably , nonlocal correlations are vital for the understanding of cuprate superconductors . but as the above aspects suggest , some of the non - fermi liquid signatures can be produced or at least aided by the presence of a van hove singularity in the vicinity of the fermi surface . this is especially important for lattices without nesting or with frustration , where nonlocal antiferromagnetic spin - fluctuations are usually suppressed and thus might be too weak to produce non - fermi liquid signatures . therefore , we argue that the interpretation solely in terms of collective modes which couple to the electronic degrees of freedom is too simplified and the influence of the saddle points in the brillouin zone has to be reconsidered . in the present work we have focused on the influence of a van hove singularity in the non - interaction density of states on the low temperature properties of the hubbard model . in order to include non - perturbative correlation effects we used the dmft to calculate the spectral function and self - energy . as impurity solvers we employed the enhanced non - crossing approximation for finite temperatures @xmath180 and the numerical renormalization group for @xmath22 . both methods yield dynamic quantities directly on the real frequency axis , which avoids the inaccuracies connected with a numerical analytic continuation of imaginary time data . the van hove singularity causes profound changes of the low - energy fermi liquid properties usually encountered within dmft for the hubbard model . for the strongly correlated metal close to the mott hubbard metal - insulator transition the imaginary part of the self - energy develops an unusual double - well structure at finite temperatures . this anomalous structure appears at temperatures on the order of the many - body scale @xmath0 which also determines its energy spread . it originates from an enhanced scattering of the low - energy quasiparticles at the saddle points in the dispersion relation associated with the van hove singularity . using a model dos , we have shown analytically that the non - fermi liquid signature in the quasiparticle self - energy is directly linked to non - analytic logarithmic contributions to the lattice scattering matrix . as a consequence the medium for the effective impurity develops a dip at the van hove singularity . in case of a logarithmically diverging non - interacting dos this dip produces a soft - gap in the effective medium at @xmath22 . at zero temperature the interacting spectral function for a square lattice exhibits a logarithmic divergence for half - filling . the system is well characterized by a generalized fermi liquid where the quasiparticle weight remains finite and the imaginary part of the self - energy vanishes at the fermi level . but the frequency dependent scattering rate increases with a non - fermi liquid exponent @xmath181 . upon doping a pinning of the van hove singularity to the fermi level is observed . its signature in the spectral function is always renormalized toward the fermi level and stays close to it up to rather large doping @xmath123 . for the square lattice with a finite next - nearest neighbor hopping @xmath23 the van hove singularity in the non - interacting dos is located below the fermi level and moved toward it upon increasing hole - doping . this trend is also observed in the anomalous maximum in the self - energy . in case of half - filling and electron doping , as well as for large hole doping , the zero - frequency quasiparticle scattering rate exhibits a low temperature dependence which is consistent with that of a two - dimensional fermi liquid . in between these fillings the decrease is much slower and follows a power law with exponent close to one . a similar signature of the van hove singularity was found in the resistivity . a fermi liquid @xmath168 dependence emerged for @xmath149 and @xmath153 , while for @xmath182 the resistivity was enhanced and a decrease with exponents less than two was observed . the findings of this work bear some implications on theories for cuprate superconductors . even thought no nonlocal fluctuations were included , we still produced qualitative features usually only obtained within more advanced theories where those correlations are incorporated . this raises the questions concerning the origin and physical mechanism behind such features and we think the role of the van hove singularity in connection with strong correlations should be further explored in the future . the author acknowledges very fruitful discussions with n. grewe , f. b. anders , t. jabben , e. jakobi and f. gttge . the author especially thanks f. b. anders for providing him with his nrg code with which the zero temperature calculations of this work were done . this work was supported by the deutsche forschungsgemeinschaft under grant no . an 275/6 - 2 . the scattering matrix can be calculated analytically via eq . ( [ eq : scatter ] ) for any @xmath183 ( omitting the arguments ) , @xmath184 \\ & \overset{\alpha=0}{= } \frac{1}{\tilde{g}}\left [ -1 + \frac{\text{li}_2(\tilde{g}w ) -\text{li}_2(-\tilde{g}w)}{2\tilde{g}w } \right ] \quad,\end{aligned}\ ] ] where @xmath185f@xmath186 is the gauss s hypergeometric function and li@xmath187 the dilogarithm . equation ( [ eq : mediumsc ] ) then yields the effective medium for @xmath55 , eq . already constitutes the final expression , while for the @xmath189 equation can be further simplified . for @xmath61 the hypergeometric function reduces to @xmath190 \quad , \end{aligned}\ ] ] leading to the effective medium displayed in eq . ( [ eq : hubtoylindos1 ] ) . in case of @xmath62 @xmath191 \quad , \end{aligned}\ ] ] yields the effective medium of eq . ( [ eq : hubtoylindos2 ] ) . assuming a fermi liquid as in eq . ( [ eq : flsig ] ) and taking the limit @xmath192 the effective media take the approximate low - energy forms , @xmath193 ( @xmath194 , @xmath22 , @xmath195 ) @xmath196 for @xmath55 the effective medium reaches zero at zero energy . for @xmath61 it saturates at a value @xmath197 but the non - analytic cusp still remains , while for @xmath62 the cusp is removed and im@xmath198 .	when a van - hove singularity is located in the vicinity of the fermi level , the electronic scattering rate acquires a non - analytic contribution . this invalidates basic assumptions of fermi liquid theory and within treatments based on perturbation theory leads to a non - fermi liquid self - energy and transport properties . such anomalies are shown to also occur in the strongly correlated metallic state within dynamical mean - field theory . we consider the hubbard model on a two - dimensional square lattice with nearest and next - nearest neighbor hopping within the single - site dynamical mean - field theory . at temperatures on the order of the low - energy scale @xmath0 an unusual maximum emerges in the imaginary part of the self - energy which is renormalized towards the fermi level for finite doping . at zero temperature this double - well structure is suppressed , but an anomalous energy dependence of the self - energy remains . for the frustrated hubbard model on the square lattice with next - nearest neighbor hopping , the presence of the van hove singularity changes the asymptotic low temperature behavior of the resistivity from a fermi liquid to non - fermi liquid dependency as function of doping . the results of this work are discussed regarding their relevance for high - temperature cuprate superconductors .
the 2df qso redshift survey ( 2qz ) aims to compile a homogeneous catalogue of @xmath7 qsos using the anglo - australian telescope ( aat ) 2-degree field facility ( 2df ) @xcite . this catalogue will constitute a factor of @xmath8 increase in numbers to a equivalent flux limit over previous data sets @xcite . the main science goal of the 2qz is to use qsos as a probe of large - scale structure in the universe over a range of scales from 1 to @xmath9 , out to @xmath10 . these measurements have the potential to help determine the fundamental cosmological parameters governing the universe . the 2qz will also significantly advance our understanding of the agn / qso phenomenon , by allowing us to carry out statistical studies of large numbers of qso spectra , and also discovering rare and extreme types of agn , for example broad - absorption - line qsos or post - starburst qsos @xcite . as well as qsos , there are a large number of other interesting sources being discovered including compact narrow - emission - line galaxies , white dwarfs and cataclysmic variables . in this paper we will concentrate on the statistical measurements of qso clustering . the basic properties of the 2qz are outlined in section [ sec : survey ] . we then discuss various measurements of large - scale structure including the correlation function ( section [ sec : xir ] ) and redshift - space distortions ( section [ sec : zspace ] ) . the 2qz currently ( september 2001 ) contains 20573 qsos below a redshift of @xmath10 . observations will be completed in january 2002 , by which time close to 25000 qsos will have been observed . a large fraction of the data is already publicly available ; the 2qz `` 10k catalogue '' @xcite contains spectra of 11000 qsos , and almost 10000 other sources . the spectra and catalogue are available electronically at http://www.2dfquasar.org . the analysis below mostly concerns this 10k catalogue . qso candidates were selected as being point sources and bluer than the stellar locus , based on broad band @xmath11 magnitudes from automatic plate measuring ( apm ) facility measurements of uk schmidt telescope ( ukst ) photographic plates . the magnitude limits are @xmath12 . the survey comprises 30 ukst fields , arranged in two @xmath13 declination strips centred in the south galactic cap ( sgc ) at @xmath14 and the north galactic cap ( ngc ) at @xmath15 with ra ranges @xmath16 to @xmath17 and @xmath18 to @xmath19 respectively . the completed survey will cover approximately 740 deg@xmath20 @xcite . spectroscopic observations have been carried out using the 2df instrument at the aat in conjunction with the 2df galaxy redshift survey ( 2dfgrs ) @xcite , as the 2qz and 2dfgrs areas cover the same regions of sky . spectroscopic data are reduced using the 2df pipeline reduction system @xcite . the identification of qso spectra and redshift estimation was carried out using the autoz code written specifically for this project . this program compares template spectra of qsos , stars and galaxies to the observed spectra . identifications are then confirmed by eye for all spectra . the mean spectroscopic completeness for the sample is 89 per cent . the spatial and redshift distribution of qsos is shown in fig [ fig : wedge ] . at redshifts @xmath21 the decline in qso numbers is due to the increased reddening of qso colours by absorption in the ly@xmath22 forest . qsos in this region are therefore missed by our colour selection . at low redshifts we will miss objects in which the host galaxy contributes significantly to the flux , due to both the extended nature of the sources and their redder colours . an estimate of the survey incompleteness due to the colour selection is given by boyle et al . ( 2000 ) @xcite . we have measured the redshift space correlation function , @xmath23 , of 2qz qsos , both averaged over the entire sample , and sub - divided into redshift or magnitude intervals . @xmath23 has been estimated assuming two representative cosmological models ; the @xmath24 einstein - de sitter model ( eds ) and a model with @xmath2 and @xmath3 ( @xmath25 ) . we use the minimum variance @xcite correlation function estimator , and details of our method can be found in croom et al . ( 2001 ) @xcite . in fig . [ fig : xi]a shows the qso correlation function for eds and @xmath25 universes averaged over @xmath26 , based on the qsos in the 10k catalogue . the amplitude and shape of @xmath23 is comparable to that of local galaxies ( fig . [ fig : xi]b ) . fitting a standard power law of the form @xmath27 we find that @xmath28 and @xmath29 for an eds cosmology . the effect of a significant cosmological constant is to increase the separation of qsos , so that in the @xmath25 model the best fit is @xmath30 and @xmath31 . both these best fit lines are shown in fig [ fig : xi]a . we then sub - divide the 2qz into five redshift intervals containing approximately equal number of qsos . the correlation function is measured in each redshift interval separately , and a power law is fit to the result ( assuming the same slope as found from the full sample ) . the resulting clustering scale lengths are shown in fig . [ fig : xi]c . the clustering of qsos is constant as a function of redshift over the entire range probed by the 2qz ( eds ) . in the @xmath25 case there is a marginal increase of clustering with increasing redshift , but a constant value can not be ruled out . making comparisions to the clustering of high redshift ly - break galaxies @xcite , at @xmath32 , we see that these also have a similar clustering strength to that of the 2qz qsos . the solid lines in fig . [ fig : xi]c denote the linear theory evolution of clustering . the data disagrees with the linear theory prediction , implying that qsos must have a redshift dependent bias factor , @xmath33 . we derive an empirical fit to the bias of the qsos assuming @xmath34 where @xmath35 is the linear growth factor ( dot - dashed lines in fig . [ fig : xi]c ) . the best fit values are @xmath36 and @xmath37 ( eds ) and @xmath38 and @xmath39 ( @xmath25 ) . a more detailed analysis is carried out by croom et al . ( 2001 ) @xcite . we lastly sub - divide our sample into 3 apparent magnitude slices with equal number of qsos in each . the measured values of @xmath40 as a function of mean apparent magnitude are shown in fig . [ fig : xi]d . there is no significant difference between the different magnitude slices , although the brightest qsos do have marginally stronger clustering . we note that sub - dividing on the basis of apparent magnitude is approximately equvalent to selection relative to @xmath41 due to the extreme luminosity evolution of qsos over the redshift range of our sample @xcite . by comparing clustering along and across the line of sight , and modelling the effects of peculiar velocities ( both linear and non - linear ) it is possible to detect the geometric distortions present if the wrong cosmological model is assumed when determining the clustering @xcite . at high redshift , the geometric distortion is particularly sensitive to the cosmological constant @xmath4 . in practice , a model which also takes into account linear infall is fit to provide a constraint in the @xmath4 vs. @xmath42 plane . the qso power spectrum from the 2qz 10k sample , measured along and across the line of sight , @xmath43 , is shown in fig . [ fig : zspace ] . unfortunately the current 2qz data only provides a joint constraint on @xmath44 and @xmath45 ( shading in fig . [ fig : zspace]b ) . infact there is also some dependence on the value of the small scale non - linear pair - wise velocity dispersion , @xmath46 . because we are looking at @xmath43 on large scales this is not a major issue , but the effects of varying the assumed velocity dispersion is shown in fig . [ fig : zspace]b . however , we can obtain a useful constraint on the cosmological world model by combining the above approach with our knowledge of the linear growth of clustering . by using the value of @xmath45 and the two - point correlation function for nearby galaxies derived from the 2dfgrs @xcite we can determine the clustering of matter at @xmath47 and for a given cosmology we can then derive the clustering of mass as a function of @xmath48 . comparison to the amplitude of qso clustering then gives the mean value of the qso bias and hence @xmath45 . this approach gives a differing relation between @xmath49 and @xmath45 ( dotted and dot - dashed lines in fig . [ fig : zspace]b ) , and thus breaks the degeneracy . the current best fit derived from this method is @xmath50 and @xmath51 . ths analysis is discussed further in outram et al . , ( 2001 ) @xcite . with the completion of the 2qz at the end of 2001 , analysis of the complete data set will produce further constraints on cosmological world models . this will include different types of analysis than those discussed here , including the gravitational lensing properties of the 2qz qsos . it is hoped that the 2qz will become a major resource for the astronomical community , particularly when the final data becomes public at the end of 2002 . adelberger k. l. , steidel c. c. , giavalisco m. , dickinson m. , pettini m. , kellogg m. , 1998 , apj , 505 , 18 alcock c. , paczyski b. , 1979 , nature , 281 , 358 bailey j. , glazebrook k. , 1999 , 2df user manual , anglo - australian observatory ballinger w. e. , peacock j. a. , heavens a. f. , 1996 , , 282 , 877 boyle b. j. , fong r. , shanks t. , peterson b. a. , 1990 , 243 , 1 boyle b. j. , shanks t. , croom s. m. , smith r. j. , miller l. , loaring n. , heymans c. , 2000 , , 317 , 1014 brotherton m. s. et al . , 1999 , apj , 520 , l87 colless m. , 1999 , in morganti r. , couch w. j. , eds . , eso / australia workshop , looking deep in the southern sky . springer - verlag , p.9 croom s. m. , 1997 , ph.d . thesis , university of durham croom s. m. , shanks t. , boyle b. j. , smith r. j. , miller l. , loaring n. , hoyle f. , 2001 , , 325 , 483 croom s. m. , smith r. j. , boyle b. j. , shanks t. , loaring n. s. , miller l. , lewis i. j. , 2001 , , 322 , l29 landy , s. d. , szalay , a. s. , 1993 , apj , 412 , 64 lewis i. j. et al . , 2001 , submitted outram p. j. , hoyle f. , shanks t. , boyle b. j. , croom s. m. , loaring n. s. , miller l. , smith r. j. , 2001 , , in press , ( astro - ph/0106012 ) peacock , j. a. et al . , 2001 , nature , 410 , 169 ratcliffe a. , shanks t. , parker q. a. , fong r. , 1998 , mnras , 296 , 173 smith r. j. , 1998 , ph.d . thesis , university of cambridge smith r. j. , croom s. m. , boyle b. j. , shanks t. , miller l. , loaring n. s. , 2001 , , submitted tucker d. l. et al . , 1997 , , 285 , l5	we present clustering results from the 2df qso redshift survey ( 2qz ) which currently contains over 20,000 qsos at @xmath0 . the two - point correlation function of qsos averaged over the entire survey ( @xmath1 ) is found to be similar to that of local galaxies . when sub - dividing the sample as a function of redshift , we find that for an einstein - de sitter universe qso clustering is constant ( in comoving coordinates ) over the entire redshift range probed by the 2qz , while in a universe with @xmath2 and @xmath3 there is a marginal increase in clustering with redshift . sub - dividing the 2qz on the basis of apparent magnitude we find only a slight difference between the clustering of qsos of different apparent brightness , with the brightest qsos having marginally stronger clustering . we have made a first measurement of the redshift space distortion of qso clustering , with the goal of determining the value of cosmological parameters ( in partcular @xmath4 ) from geometric distortions . the current data do not allow us to discriminate between models , however , in combination with constraints from the evolution of mass clustering we find @xmath5 and @xmath6 . the full 2qz data set will provide further cosmological constraints .

Subsets and Splits